INFORMATION TO USERS
The most advanced technology has been used to photo graph and reproduce this manuscript from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are re produced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book. These are also available as one exposure on a standard 35mm slide or as a 17" x 23" black and white photographic print for an additional charge.
Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.
UMI University Microfilms International A Bell & Howell Information Company 300 North Zeeb Road. Ann Arbor, Ml 48106-1346 USA 313/761-4700 800/521-0600
Order Number 8918677
Reliability and validity of a human figure drawing screening measure to identify emotional or behavioral disorders in children
McNeish, Timothy Jay, Ph.D.
The Ohio State University, 1989
Copyright ©1989 by McNeish, Timothy Jay. All rights reserved.
UMI 300 N. Zeeb Rd. Ann Arbor, Ml 48106
RELIABILITY AND VALIDITY OF A HUMAN FIGURE DRAWING SCREENING MEASURE
TO IDENTIFY EMOTIONAL OR BEHAVIORAL DISORDERS IN CHILDREN
DISSERTATION
Presented in Partial Fulfillment of the Requirements for
the Degree of Doctor of Philosophy in the Graduate
School of the (Miio State University
By
Timothy Jay McNeish, B.S., MA
**********
The Ohio State University
1989
Dissertation Committee; Approved by
H. Leland, Ph. D.
P. Post, Ph. D.
W. B. Walsh, Ph.D. reri,Ph. D.
Adviser
Department of Educational Services
and Research Copyright by
Timothy Jay McNeish
1989 Writing a book is ai adventure; to begin with it is a toy and an anusement, then it becomes a master, and then it becomes a lyrait ; aid the last phaœ is just as you are about to be reconciled to your servitude - you kill the monster and fling him... to the public.
- Sir Winston Churchill To My Parents ACKNOWLEDGMENTS
My sincere appreciation goes out to the hundreds of students who participated in this
project, end to the administrators æd staff members at each facility who tolerated my
presence and æsisted in the data collection. I am indebted to my dissertation committee
members, Dr. Jack Naglieri, Dr. W. Bruce Walsh, Dr. Henry Leland, and Dr. Phyllis Post for their helpful suggestions Ëx)ut this project and for their teachings and contributions throughout the doctoral program. I am especially thankful for the encouragement aid expertise offered by Dr. Naglieri who chaired my dissertation conmittee and served as my adviser. His friendship and scholarship have been invaluable.
I am also indebted to those who served as research assistants for this stud/: Judy
McNeish, Betty Carver, Barbara Bell, and Kathy Renard. The precision and effort with which they worked greatly contributed to the project's completion. 1 am also very appreciative of the frequent words of encouragement shared by these friends, and by many other friends and family members.
This dissertation is dedicated to my parents, John and Donna McNeish. Throuÿwut my development they have encouraged every goal, applauded every accomplishment, aid accepted every shortcoming. They have provided as fine of models as a son could wish for and any positive contribution I might mdce as a person or a professional is a direct reflection of them.
For all of this, I express my sincere gratitude
The contributions of my wife, Judy, toward the completion of this project aid of the doctoral deyee have been beyond measure. She made innumerable personal sacrifices, showed tolerance of my absence from home, and displayed patience with my work pace and project iii preoccupations. For her faith in my abilities and the love, support, and encouragement she offered at every stage, I am deeply and forever ^preciative. I recognize her contributions with love and pride.
iv VITA
June 6,1957 ...... Born - Zanesville, Ohio
1979 ...... B. S., The Ohio State University, Columbus, Ohio
1979-1981 ...... Student Personnel Assistent The Ohio State University, Columbus, Ohio
198 1...... M. A, The Ohio State University, Columbus, Ohio
1981-1984 ...... Psychology Assistait Moundbuilders Guidance Center Newffl'k,Ohio
198 2 ...... Adjunct Faculty, Park College Newark, Ohio
1984-1986 ...... Graduate Research Associate The Ohio State University, Columbus, Ohio 1985 ...... Graduate Research Associate Ctrarles Merrill Publishing Oompaiy Columbus, Ohio
1986-198 7 ...... Psycholog/Intern Moundbuilders Guidance Center Newark, Ohio
1987-Presen t ...... Psychology Intern Psychological Associates Newark, Ohio
PUBLICATIONS
McNeish, I. J. & Price, P. S. ( 1986). Criterion-related validity of the Matrix Analogies Test - Short Form in the assessment of the hearing impaired [Summary]. Proceedinos of the 18th Annual Convention of tl% National Association of School Psvcholooists. pp. 281-282.
FIELDS OF STUDY
Major Field: Counseling
Studies in Psychodiagnostic and Developmental Assessment, Neuropsychology, Counseling Psychology, Cognitive Psychology, and Social Psychology
Vi TABLE OF CONTENTS
DEDICATION...... ii
ACKN0WLEW3MENTS...... iii
VITA...... V
LIST OF TABLES...... viii
LIST OF FIGURES...... x
CHAPTER PAGE
I. INTRODUCTION...... 1
II. REVIEW OF THE LITERATURE...... II
III. METHODOLOGY...... 45
IV. RESULTS...... 61
V. DISCUSSION...... 81
APPENDICES
A. Administration Instructions ...... 90
B. Preliminary Scorit^ Manual ...... 91
C. Final Scoring Manual ...... 105
D. Slanted Figure Scoring Template...... 116
REFERENCES...... ! I?
v ii LIST OF TABLES
TABLE PAGE
1. Empirical Findings for Validity of HFD Structural or Formal Items ...... 34
2. Empirical Findings for Validity of HFD Content Items ...... 35
3. Normal Group Characteristics by Region and Age ...... 46
4. Normal Group Characteristics by Race and Age ...... 47
5. Exceptional Group CM'acteristics by Collection Site and Age ...... 50
6. Exertional Group Characteristics by Race and Age ...... 51
7. Percentage of Item-Agreement by Rate^ on Competency Exam...... 62
8. Cut-off Scores for Structural Items on the Preliminary Scale...... 63
9. Frequency of Occurrence for Each Prelimirwy Scale Item Among Normal Subjects (n = 186) on the Man, Woman, æd Self Drawings, and the Mean Occurrence Across All Three Drawings ...... 64
vin 10. Percentage of Intra-rater and Inter-rater Item-Agreranent for 25 Drawings ...... 68
11. Preliminary Scale Item-Total Correlations...... ?2
12. Means, Standard Deviations, and 1-Test for HFD: FDS Group Differences...... 75
13. Cirrelatlon Matrix for HFD: FDS %ores Between Each individual Drawing, and Between Each irehvidual Drawing and the Remaining Two Drawings ( hbrmals, n = 186) ...... 76
14. Gorrelatiwi Matrix for HFD: FDS Scores Between Each Individual Drawing, and Between Each Iralivldual Drawing md the Remaining Two Drawings ( Exceptionals, n = 92) ...... 76
15. Means, Standffl^ Deviations, and F-Test for Total Number of Items Present per Drawing (Normals, n= 186) ...... 77
16. Means, Standard Deviations, and F-Test for Total Number of Items Present per Drawing ( Exceptionals, n = 92) ...... 76
17. Means, Standa^ Deviations, and F-Tests for HFD: FDS Race/Regional Differences ...... 79
IX LIST OF FIGURES
FIGURE PAGE
1. Am Length Measurement ( Preliminary Scoring Manual) ...... 92
2. Leg Length Measurement (Preliminary Scoring Manual) ...... 93
3. Hand Heiÿit Measurement (Preliminary Scoring Manual) ...... 93
4. Hand Width Measurement (Preliminary Scoring Manual) ...... 94
5. Foot Length Measurement (Prelim inay Scoring Mmual) ...... 94
6. Width of No'rowest Portion of Tora) Measurement ( Preliminar/ Scoring Manual) ...... 95
7. Stance Width Measurement (Preliminary Scoring Manual) ...... 95
8. Head Width M easuranenKPrelim in^ Scoring Manual) ...... 96
9. Head Heiÿit Measurement (Preliminary Scoring Manual) ...... 96
10. Leg Length Measurement ( Final Scoring Mmual ) ...... ! 06
11. Foot Length Measurement (Final Scoring Manual) ...... 107
12. Width of Narrowest Portion of Torso Measurement (Final Scoring Manual) ...... 108
X 13. Stance Width Measurement ( Final Scoring Manual) ...... 108
14. Head Width Measurement (Final Scoring Manual) ...... 109
15. Head Height Measurement (Final Scoring Manual) ...... 110
16. Arm Length Measurement (Final Scoring Manual) ...... 11 i
17. Hand Heiÿïî McuSüféfnent ( Final Scoring Manual) ...... 1 i 2
18. Hand Width Measurement (Final Scoring Manual) ...... 112
XI Chapter I
Introduction
This study examined the relationship between chllA ^'s Human Figure Drawings
( HFDs) and their emotional atgustment or psydwloglcal well-being. The study was based on the comparison of HFDs produced by normal chllAwi and those rendered by children who have been Identified as having m emotional or behavioral disorder. This first chapter presents the background of the study Including a description of the technique and Its early development The theory underlying the HFD technique and the relevant reseatrfi are briefly reviewed. The chapter concludes by describing the stucVs significance and specifying the problem to be
Investigated.
Description of the Technique
The HFD technique has two principle applications (Koppltz, 1968, p. I). It Is used both as a test of intellectual or developmental maturity and as a projective personality assessment technique. Thus, throuÿi this brief method of assessment a clinician can lev'n valuable
Information ^xtut a subject’s Intellectual ability gjd emotional acQustment or personality characteristics. The technique Is administered by examiner to an Individual subject or a group of subjects In a period of five to fifteen minutes (Koppltz, 1968; Settler, 1988). The
Instructions vary slightly depending upon the Intended use of the drawings ( I.e., as a projective technique or as an estimate of Intellectual capacity), but the essential components
the sane. The subject Is Instructed to draw three human figures, one of each sex, and a self-drawing. Each figure Is drawn on a separate, unlined sheet of eight and one-half by eleven inch white paper. Upon completion, the examiner rates each drawing using a multi
dimensional procedure in order to make a determination about the subject's level of
intellectual development, emotional adjustment, and/or personality characteristics.
When used to assess intellectual maturity (or developmental level), the drawings are
scored based on the inclusion of bod/ parts, clothing details, proportion, perspective, aid
similar features (Anastesi, î 988). For example, Draw-a-Man scale item 30 of the
Goodenough-Harris Drawing Test (0HDT) (Harris, 1963) reads:
Arms Present - Any method of representation clearly Intended to indicate arms. Fingers alone are not sufficient, but the point Is credited if any space is left between the base of the fingers and that part of the bod/to which th e/attach ed . The number of arms must be correct, except in profile drawings when only one a-m may score, (p. 254)
Raw scores are totaled and converted to standad scores and compared to a normative sanple.
These scores serve as an estimate of IQ and are used as one component of the overall
psychological assessment battery.
When used as a projective assessment technique a somewhat different set of scoring
rules is used The drawings are examined for structural and formal aspects ( size, placement
on page, symmetry, perspective, line quality, etc.), overall mood or tone (facial expression
and posture), content (presenceof various bod/ parts), and conflict indicators (differential
treatment of bod/ parts throuÿi omissions, breaks, dimming or reinforcing, erasures, or
shading) (Hammer, 1958;Koppit2, 1968,1984; liachover, 1949). Hypotheses, based
largely on clinical æcperience and/or a psychod/namic approach to drawings, are then generated about potential behavioral attributes or areas of intrapersonal or interpersonal conflict the subjsct mey be experiencing. These hypothesæ are typically viewed in ttie context of other clinical assessment data to formulate a description of the subject's personality and
emotional adjustment
Theoretical Bases
The use of HFDs 8S8measure of intellectual capacity is grounded in a theory which describes intelligence as the ability to form concepts of an abstract character (Harris,
1963). According to Sattler ( 1988), "the abilities involved in forming these concepts are
(a) perception [discrimination of likenesses and differences], (b) abstraction [classification of objects], md (c) generalization [assigning newly experienced objects to the correct class]"
( p. 311 ). Sattler suggested these drawings reflect the complexity of tfie subject's concept formation ability. Sattier also explained that the human figure drawing is more ^propriété for this purpose than drawings of mother type because "...it is the most familiar and meaningful figure for the diild" ( p. 311 ).
The interpretation of HFD s as reflective of personality characteristics or emotional adjustment is couched in the projective hypothesis or projective method Frank ( 1948) described this method as "studying the perstmality by confronting the subject with a situation to which he will respond according to what the situation means to him and how he feels when so responding" ( p. 46). This process is referred to as apperception, or the subjective interpretation of a perc^tion meaningful to the person making i t The apperceptive process has be defined by Lelmd ( 1983) as "the manner in which individuals deal with their inrer promptings md the demands of society as filtered through other aspects of their inner promptings" (p. 147). This becomes a rather circular process and Leland further noted that one's response, or behavior, is based on this circularity. According to Frank ( 1948), "the essential feature of a projective technique is that it evokes from the subject what is in various ways ® geographical, cultural, and social environments" ( Rabin, 1981, p. 10). Historical Develmment The relationship between the completeness or detail of children's drawings and their level of development was observed and documented as early as the late 19th century ( Oooke, 1885; Ricci, 1887) and support for this relationship has accumulated throuÿi the present time(Ooodmouÿ), 1926; Harris, 1963,Koppit2,1968,1984; Naglieri, 1988). Burt (1921) presented evidence of developmental progress and delay in childrm's (Sawings and included a drawing of a man as one of a number of "Scholastic and Mental Tests." Similarly, Goodenouÿi ( 1926) published findings in her text Measurment of Intelligence bv Drawirxs that chilc^en's drawings relate to intellectual maturity. This resulted in the development of the Draw-A-Mai Test (DAM) (Goodenouÿi, 1926) which was readily accepted and became widely used as a measure of children's intellectual Aility ( Koppitz, 1968). Goodenough's original test was used for nearly 40 years before being expandal and refined by Harris ( 1963) (The Goodenough-Ho'ris Drawing Test [GHDTl). Harris' revision added a drawing of a woman and a self-tfr-awing to the test, provided a nationally representative norm group, and broadened the age-range to which the test could be applied. Recently, Naglieri ( 1988) published a similar approach to rating children's intelligence based on their HFDs (Draw-A-Person: A Quantitative Scoring System [DAP-QSSJ) which was responsive to criticisms leveled(e.g., Sattler, 1982,1988)at theOoodenough-Ho'rissystem. Specifically, Naglieri modernized the scoring sidelines for the womoi drawing, attempted to reduce the ambiguity of the ovo'all scoring criteria, provided a current nationally representative standoxlization sample, and refined the norms into half-year and quarter-yeer intervals to increase the precision of the scale. Concurrent with these advances regaling HFDs as a measure of intellectual or developmental level, an alternate application of the technique was developing. Drawing differences on HFDs which were seemingly unrelated to intellectual or developmental level became apparent by the early 1920's (Jones & Thomas, 1961, p. 214). Regaling these differences, Ooodenough ( 1926) stated: "...the/ probably carry profound mealing, had we but the wisdom to understoid it" ( p. 60). This concept, that HFD s revealed more dwut the drawer than intellectual ability, reappeared in Machover's ( 1949) Personality Projection in the Drawing of the Human Figure, liachover, basedon her clinical work with adolescents and adults, suggested: "the human figure drawn by an individual who is directed to 'drwt a person' relates intimately to the impulses, anxieties, conflicts, and compensations characteristic of that individual" ( p. 35). More specifically, Machover believed particular signs in these tfr'awii'qs were directly indicative of specified psychological conflicts. For example, "speo^ or talon-like fingers" ( p. 64) were viewed as pathognomonic of aggressive impulses and "excessive emphasis on the eyes" (p. 16) was believed to be associated with paranoid ideation. Machover's book has become the most widely quoted book in the field and has equaled Doodenough's book in significance and influence ( Koppitz, 1968, p. 2). Her ideas have been vigorously supported through the years by Hammer (1958,1959,1969) and catalogs designed to aid in this method of HFD into-pretation have been developed by Urboi ( 1963) and Ogden ( 1977). The majority of reseoxh and writing regarding the use of HFDs as a projective assessment technique has focused on adult populations ( Falk, 1981), however, similar principles and theory have beoi ^plied to the HFDs of children. Koppitz ( 1968) developed the first refined scoring system to evaluate the HFDs of children (Klepsch & Logie, 1982, pp. 18-19). Her approach combined developmental and projective aspects of the drawings and thus, included two sets of items. She considered both aspects as critically important for every child's drawing. Her Developmental Items consisted of those derived from tte Goodenough-Harris ( 1963) scoring system and from Koppitz' own experience working with children. The Emotional Indicators were derived fron the work of Machover ( 1949) and Hammer ( 1958) and, again, her own clinical experience. While the Development! Items were defined in a manner similar to previous HFD developmental scoring systems (i.e., items occur relatively infrequently in t e drawings of young children and occur with increasing frequency as the age of the children incraaæs), the Emotional Indicators were more strictly defined and interpreted than in traditional projective HFD approaches. For example, the selection of items as Emotional Indicators was based on three specific criteria First, items had to demonstrate clinical validity, that is, be able to differentiate between the HFDs of children with emotional problems and those without Second, each item had to occur infrequently in the HFDs of normal children. And third, an item's frequency of occurrence could not increase solely on the basis of the children's increase in age. Koppitz' systematic item selection process and emphasis on both developmental and projective aspects of HFDs reflects a fuller appreciation of the technique's versatility and a stronger commitment to demonstrating its validity than past efforts. Her contributions served as a foundation and rough framework for the current study. Related Research The HFD technique has bem widely investigated both as an intellectual measure for children and as a projective p^'sonality measure for adults and, to a lesser degree, children. Buros ( 1978) ranked the Goodenouÿi-Harris Drawing Test ( Harris, 1963) and the Machover ( 1949) Draw A Person Test as the 29th and 40th most frequently researched psychological instruments, respectively. While reliability and validity studies have consistently supported the technique as a screwing measure for childhood Intellectual development ( Harris, 1963 ; Naglieri, 1988; Settler, 1988), research on its use as an indicator of personality characteristics or emotional adjustment has been conflicting. Numerous case studies appear in the literature which attest to the technique's usefulness in isobted clinical situations, yet experimental research has offered little to no support for the technique's validity as an indicator of emotional or behavioral attributes (Maloney and Olasser, 1982, p. 183). Comprehensive reviews of the HFD literature wereconcfcicted by Swensen ( 1957,1968), Roback ( 1968), and Kahili ( 1984) and each reviewer concluded Machover's original hypothesis lacks empirical support Koppitz' system of Emotional Indicators on children's drawings has similarly failed the test of diagnostic validity ( Dieffenbach, 1977; Mortensen, 1984; Pihl & Nimrod, 1976; Snyder & Gaston, 1970). Several important findirgs aid recommendations teve resulted from the investigations to date which are relevant to future research when HFDs are used as an indicator of emotional or behavioral problems. First, there is a need for refinement and increased objectivity among HFD scoring items as the reliability with which items are scored has been unsatisfactory (Roback, 1968; Swensen, 1957,1968). Second, items which have been theoretically associated with disturbance appear in the HFDs of normal individuals as well as the HFDs of disturbed individuals (Koppitz, 1968;Naumberg, 1953 [cited in Falk, 1981]; Ogdon, 1977). Based on this observation, several investigators have pointed out it is tlie number of items present in a drawing which differentiates normal and maledj^^tad subjects (Koppitz, 1968; Fteumberg, 1953 [cited in Falk, 1981]; Ogdon, 1977). Thus, the clinician must be thoroughly ( 001110* with the drawings of normal subjects before using the technique to assess disturbance. Third, even thoüÿi the hypothesized one-to-one relationship between individual items and specified psychological conflicts lacks empirical support, global or overall ratings of HFDs have been found to significantly relate to gross levels of maladjustment ( Kahili, 1984;R(t8ck, 1968; Swensen, 1957,1968). Therefore, HFD interpretation should be based on global aspects of the drawings and the technique should be applied as a screening measure for gross levels of maladjustment rather than as an indicator of a particular conflict anomaly. Fourth, a tendency for clinicians to use the HFD technique as either a measure of intellect w a projective tool has been observed. Both aspects of the HFD are critically important for every drawing and failure to recognize this harshly limits the technique’s value (Adler, 1970; Koppitz, 1968). Last, it has been argued that HFD projective items reflect 0 » same factor as those items used to assess intellectual or developmental level. Factor analytic studies ( Nichols ScStrumpfer, 1962; Adler, 1970) have revealed that a single factor, cognitive maturity, accounted for over half of the common variance of HFDs. However, three less powerful factors were also evident which, according to Adler, "... may bear some relationship to significant personality or diagnostic variables" (p. 57). Simificance of the Stud/ The importance of this study is threefold. First, despite the lack of support through quantitative analyses regarding the HFD technique's effectiveness as a projective assessment technique, its use for this purpose remains widespread. Machover's ( 1949) Draw-A-Person Test ( DAP) ranged in position from second to eleventh among the most frequently used psychological instruments between 1959 and 1984, placing eighth in the most recent survey (Sundberg, 1961 ; Wi1dman& WilAnan, 1967; Lubin, Wallis, & Paine, 1971; Brown & McGuire, 1976; Wade & B ^er, 1977; Lubin, Larsen, & Matarazzo, 1984). Thus, placement or diagnostic decisions are made about individuals based, at least in part, on an assessment technique whose validity is the subject of frequent debate. Second, it has been argued that the failure to quantitatively document this technique's validity is a function of the method(s) of investigation used thus far rather than the technique itself. Roback ( 1968, p. 16) pointed to "a paucity of quality reseaxh" on the technique and called for "standardized and validated scales for estimating personality adjustment from figure drawings." Criticism has also t»en leveled regarding the attempt of many studies to deduce psychiatric diagnoses based solely on the HFD technique. Such a practice runs counter to the projective hypothesis on which the technique Is based ( Falk, 1981; Kahili, 1984). Third, even though children are natural subjects for W s type of research they are grossly under-represented in HFD studies. Falk (1981) noted over three-quarters of the studies reviewed by Roback ( 1968) and Swensen ( 1957, 1968) used only adults as subjects, and Kahili ( 1984) limited her review to studies of adolescents and adults. Problem Statement The general question addressed by this study was this: Can a a»le for scoring HFDs prove useful as a screening device for emotional malat^ustment and/or maladaptive personality traits in children? This general question subsumed the following specific Issues. The first three issues related to developing Items for such a scale and the remaining three issues adck'essed evaluating the usefulness of such a scale. Question one. Which of the HFD items reported In the literature as indicative of emotional or behavioral disturbance are rare or unusual ( occurring with less thaï 1 S% frequency) in the drawings of normal subjects? Question two. Which of the HFD items reported in the literature as indicative of emotional or behavioral disturbance can demonstrate an acceptable level of rater agreement (80%) under intra-rater and inter-rater conditions? Question three. To what dagree is each item reported In the literature as indicative of emotional or behsvicral disturbmce related to this total group of Items? Quætion four. When these Itans are combined to form a scale, cai this scale demonstrate satisfactory intra-rater and inter-rater reliability coefficients? 10 ftjestlon five. When these items are combined to form a scale, can this scale be shown to diverge from measured intellectual ability by demonstrating low correlations with standWzed measures of intelligence? Question six. When these items are combined to form a scale, can this scale differentiate between a group of children who are placed in special settings for treatment of emotional or behavioral disorders and a group of chilcHn are enrolled in reguly classrooms? Chapter II Review of the Literature The present stud/ builds on an alread/ voluminous bod/ of research on the Human Figure Drawing ( HFD) technique. The research literature clustw^ around two basic areas which reflect the technique's two principle applications. This review is organized around these areas, the first addreœing the technique's usefulness as a measure of intellectual or developmental level for children and the second, wamining its usefulness as a projective personality assessment tool fcr adults aid children. Because the usefulness of any formal test or assessment technique is largely dependent upon its technical qualities, this review focuses on such qualities for each application of the HFD technique. Walsh and Betz ( 1985) suggest the two most basic components of test quality are reliability, or "the extent to which we measuring some attribute in a systematic and therefore repeatable way" ( p. 47) and validity, or “the extent to which the test we're using actually measures the characteristic or dimension we intKKi to measure" ( p. 56). The evaluatiwi of these basic components serves to ancfiw' Uiis review. Additional HFD findings are also presented as they relate to the present stud/. HFD's as 8 Measure of Intellectual Development Children's (Sawings first received serious attention during the latter portion of the nineteenth century ( Klepsch and Logie, 1982). Early papers on the topic focused on the relationship between these drawings and the level of development of the children who produced them. Corrado Ricci ( 1887), a renowned art critic, was among the first to publish sudi a paper. Other pioneers in the stud/of children’s drawings included Sully (1895), Partridge 11 12 ( 1902), Kerschensleiner ( 1905), Levinstein ( 1905), Katzaroff ( 1910), and Luquet (1913). In 1921, Cyril Burt used the drawing of a man as one of a number of "Mental and Scholastic Tests." He noted that while the drawings were positively correlated with children's intelligence, they were not as highly correlated with intelligence as were tests of reeding, writing, spelling, or arithmetic (cited in Mortensen, 1984). Despite this finding Burt viewed the drawings as having the advantags of being less dependent on learned skills and, therefore, still a valuable reflection of childhood intellectual development Another early attempt at measuring childhood intelligence through Human Figure Drawings was made in France, by Fay ( 1923). Fay directed children to draw “A lady takes a walk and it rains" (cited in Mortensen, 1984, p. 45). With this ^proach the subjects' work was limited to five minutes and the (Sawings were scored according to the number of details present This technique was revised by Rey in 1946, but it has been rarely used outside of FrCTch-speaking countries. Similar efforts tow^d better understanding the relationship between children's Humm F i^re Drawings ( HFDs) and intellectual development continued to emerge during this period. Goodenough's ( 1926) publication of the Draw-A-Man Test (DAM) was a landknark development on the topic and served as a cornerstone for future wwk. This test was widely used in its original version for neæly 40 years. It was revised and carefully standardized by Harris in 1963 (Ooodenough - Harris Drawing Test [^D T]), and restandardized by Harris and Roberts in 1972. Other recent advances in the use of HFDs as a measure of intellectual maturity include the work of Koppitz ( 1968) who described 30 HFD Developmental Items which relate to intellectual development, and that of Naglieri ( 1988) who developed the Draw A Person Test A Quantitative Soring System ( DAP: QSS). Through the yes^ similar methods of evaluating children's HFDs have also been included as items or subtests in more comprehensive tests of intellectual ability, for example, the Draw-A-Child subtest on the 13 McCarthy Scales of Children's Abilities (McCarthy, 1972). An important, common pathway is shared t essentially grew from studies of normal development in children as deleted in their figure drawings. The focus was on the common and general traits observed in these drawings which were produced by lerge youps of normal children. This process of development is a sharp contrast to that associated with the projective application of HFDs, as will be seen later. A vast collection of studies exists which eddress the usefulness of HFDs as a measure of intellectual development Although the earliest of the above mentioned tests and techniques were not formally evaluated, empirical findings do date back to the work of Ooodenouÿ) ( 1926). Overall, the evidence has tended to support the technique's usefulness as a screening measure of cognitive ability ( Settler, 1988). Pertinent findings follow. Reliability. Several types of reliability have been studied regarding the HFD technique when it was applied as a measure intellectual maturity. The consistency with which a single rater was able tc æore a particular drawing on separate occasions (intra-rater reliability) tended to be very high. Intra-rater reliability coefficients in the 90s were observed for both the DAM (McCarthy, 1944) and the 6HDT (Dunn, 1967,-Oottling, 1985; Levy, 1971;Yater, Barclay, & McOilligan, 1969). Similarly, Oottling (1985) reported intra-rater reliability coefficients in the 90s for the DAP: QSS. The consistency with which two separate raters were Ëile to acre a particular (Rawing ( inter-rater reliability) also tended to be very high. Inter-rater reliability coefficients in the 80s and .90s were found for the DAM (McCarthy, 1944; Williams, 1935) and the GHDT (Settler,1982). Oottling (1985) found inter-rater reliability coefficients in the .90s for the DAP: QSS. While inter-rater reliability coefficients were rest presented in Koppitz' ( 1968) o'igiral work with HFD Developmental items, her text noted that she end another psychologist reached 95 percent agreement when scoring 25 protocols with these 30 items. The available data has consistently demonstrated 14 excellent scorer reliability for each of these tests or scoring systems. Test scores for a given individual across two HFD administrations separated by a specified time interval (test-retest reliability) have tended to be quite stable. Test-retest relidJility coefficients from .68 to .91 were observed for the DAM with intervals between administrations ranging from one to six weeks (Brill, 1935; McCarthy, 1944; McCurdy, 1947; Smith, 1937;Yepsen, 1929). Similar evidence has been presented for this aspect of the GHDT with coefficients falling primarily in the 50s through .80s with intervals between administrations ranging from three weeks to seven months (Sattler, 1982; Oottling, 1985). Also, Naglieri ( 1988) obtained a median test-retest correlation coefficient of .73 for the DAP: QSS when testing and retesting 112subjects from grades one through seven witha four-week intsrval between administrations. Thus, studies have uniformly shown a substantial to very strong association between scores for a given individual when tested and retested with intervals between ackninistrations ranging from one week to seven months. Two additional types of reliability have received somewhat less attention in the HFD literature, yet merit consideration when evaluating a method of assessment like the HFD technique. First, alternate forms reliAility assesses the degree to which two different fwrns of the test yield similar results" (Walsh & Betz, 1985, p. 50). Harris ( 1963) recommended that the separate drawings be regarded as alternate forms and that, for yeater reliability, the mean of the forms' standard scores be used to rate intellectual developmait This comparison of drawings, one to another, is one method to help determine the extent to which the drawings reflect some common dimension and are measuring that dimension in a relatively consistent manner. Second, internal consistency reliability addresses the extent to which each item on the test measures the same dimension as the other items on the test, or the test's homogeneity. Homogeneous items suggest a test is unidimensional with regard to what it measures while heterogeneity among items reflects a test which is multidimensional. 15 Alternate forms reliability coefficients for the 6HDT have ranged from .72 to .90 (Saltier, 1982). This type of reliability has not yet been examined for the DAP; QSS and it is not applicable to the DAM or Koppitz' Developmental Items as each of these tests require only a single HFD. However, Ooodenough ( 1926, p. 48) reported a split-half reliability coefficient of .77 for her DAM. Split-half reliability is conceptually similar to alternate forms reliability, Uius her finding was consistent with the results of the alternate forms studies. The only test for which information was available regarding internal consistency reliability is the DAP: QSS. Naglieri ( 1988) presented internal consistency (Cronbach's alplra) coefficients ranging from .83 to .89 ( median coefficient of .86) for the total test and from .56 to .78 (median coefficient of .70) for the individual drawings (man, woman, and self). The available data has revealed a very strong association between alternste forms of the GHDT and very hiÿi consistency anong DAP: QSS items. Therefore, the HFD tests studied appear homogeneous in nature, that is, they reflect some common dimension and appear to measure that dimension in a relatively consistent manner. Validity. Studies designed to evaluate the validity of the HFD technique when it was used to æsess the intellectual or developmental level of children have also been supportive. Two types of validity are of principle interest regarding HFDs: construct validity and criterion- related validity. Construct validity focuses on the deyee to which a test demonstrates a relationship with the theoretical construct or attribute it was desigied to measure. For example, since intellectual ability is expected to increase with age during childnod, scores on a valid test of intelligence should likewise show such an increase (Anastasi, 1988, p. 153). Criterion-related validity is the dagree to which a test of an attribute demonstrates an association with some independent or external indicator of the same attribute ( Walsh & Betz, 1985, p. 58). Predictive validity and concurrent validity are specific types of criterion- related validity. Predictive validity addresses the issue of how wall a present test score 16 predicts future sWtus on the criterion verieble, while concurrent velidity exemines the relationship between present test scores and present status on the criterion variable (Walsh & Betz. 1985, p. 59). Construct validity has been evaluated at the individual item level as well as the total raw score level for HFD scoring systems. Harris ( 1963) reported that each of the 51 DAM items "conformed fairly well" (pp. 69-70) to validity requirements which included “( 1 ) a regular and ( 2) fairly rapid increase in the percentage of children succeeding with the point at successive ages." Similar findings were presented for the 6HDT items (Harris, 1963, pp. 73-77), «id Koppitz (1968, p. 21) provided evidence that each of her Developmental Items increased in frequency of occurrence at successive age levels. The meai number of total raw score points earned on HFDs has also been shown to increase with an increase in age (Gottling, 1985; Koppitz, 1968; Naglieri, 1988). Thus, both individual items «id total raw scores f(r these HFD techniques appear related to the construct of intellectual or developmental level. According to Mortensen ( 1984, p. 58) the most common method of investigating HFD validity has been through concurrent validity studies. For example, studies by McElwee, ( 1932), Vane and Kessler ( 1964), Williams ( 1935), and Yepsen ( 1929) have shown a substantial to very strong relationship (correlation coefficients ranging from .53 to .72) betwe«i the DAM and the Stanford-Binet Intelligence Scale (Terman, 1916). Investigators have also observed consistsitly positive correlstiKis ranging in magnitude from low to very strong between the DAM and other measures of intelligence including the Thurstone Primary Mental Abilities Test (Thurstone, 1938) (Ansbacher, 1952; Harris, 1963), the Wechsler Intelligence Scale for Children (Wise) (Wechsler, 1949) (Ellis, 1953;Hanvik, 1953; Rottersman, 1950), and the Wechsler Adult Intelligence Scale (WAIS) (Wechsler, 1955) (Berdie, 1945; Gunzburg, 1955). Sattler ( 1982, pp. 249-250) summarized the concurrent validity (feta for the GHDT by noting a similar range of correlations with various 17 other tests of intelligence. Koppitz ( 1968, p. 31) presented correlation coefficients ranging from .55 to .63 between her HFD Developmental Items and the Stanford-Binet and from .45 to .80 between these Items and the WiSC. The DAP: QSS total test score correlated significantly (.31) with the Matrix Analogies Test - Short Form ( MAT - SF) ( Naglieri, 1985) for a sample of 594 students in grades kindergarten through three and (.27) far 1,328 students in grades four through twelve (Naglieri, 1988, p. 19). Similarly, significant correlations were observed between the DAP: QSS and the Reading and Math sections of the Multilevel Academic Survey Test (Howell, Zucker, & Morehead, 1985) (.24 and .21, respectively) for a sample of 1,328 students in yades four through twelve. Additionally, with the exception of Koppitz" Developmental items, each of the scoring systems has been shown to correlate significantly with one another. Harris ( 1963, p. 99) reported correlation coefficients ranging from .91 to .98 (median r = .97) between the DAM and the GHDT. Similarly, Naglieri (1988, p. 18) obtained a correlation coefficient of .87 between the DAP: QSS and the GHDT total scores. To summarize the concurrent validity date on HFDs, the technique has shown consistently positive correlations with other measures of intelligence with the magnitude of these correlations ranging from low (. I O's and .20‘s) to very strong (.70s end .80s). Reviewing such results led Sattler ( 1988, p. 313) to refer to HFD tests as "a relatively poor predictor of scores on other intelligence tests," but as "an acceptable screening instrument for use as a nonverbal measure of cognitive ability." The varW lity in the relationship between HFD tests and other measures of intellectual level poses a complex cpjestion. The following action examines possible explanations for this observed variability through exploring tt% research on factors which ma/ have influence performance on the technique when it was used as a measure of intellectual development 18 Oonfoundino variables. The impact of a number of extraneous factors has been invesllgalBd In an attempt to identity what factors, other than intellectual development, most strongly influence a subject's performance on HFD tests of intellectual maturity. These factors have included both person-variables (ag., characteristics of the subjects themselves including age, sex, race, and ps/choiogical ac^ustment) and situation-variables (ag., the subjects' cultural background and socioeconomic status, the subjects' artistic ability, and the impact of different types of examiners administering the HFD procedure). The available evidence regarding the effect of both types of variables on HFD performance is reviewed below. Evidence has been cited by Anastasi ( 1988) and Harris ( 1962) regarding the effects of age on HFD performance. Aside from the increase in total raw score points at successive age levels noted earlier as evidence of construct validity, HFDs teve appeared to measure different specific functions at different ages. For example, DAM test scores for fourth-grade children correlated highest with tests of reasoning, spatial aptitude, and perceptual accuracy, while for kindergarten children the DAM test correlated higher with numerical aptitude and lower with perceptual speed and accuracy than it did with the fourth graders (Ansbacher, 1952). Fairly consistent differences regarding the effects of sex of the subject on HFD performance have been observed. Harris ( 1963) and Koppitz ( 1968) both noted that, in the primary grades, the HFDs of girls tended to be superior to those of boys with respect to the level of quality and detail. This difference reportedly diminished as the children reached age eight or nine and, in fact, boys' drawings then became superior to those of girls ( Koppitz, 1968). Such sex differences required separate normative data being established for the interpretation of the GHDT and Koppitz' Developmental Items. Naglieri ( 1988) also found sex differences in the DAP: QSS stendardization deta Observed differences were statistically significant with girls earning higher scores than boys (fi < .00 Î ), yet the actual differences were small enouÿi ( less than one-fifth of the ^ of the test) that the generation of separate 19 norms was deemed unnecessary. Inconsistencies were evidenced in the literature regarding the impact that racial and cultural differences among subjects has had on HFD performmce. Significant differences were found between the GHDT scores of Mexican and America» children ( Laosa, Swartz, & Diaz- Guerrero, 1974). Also, significant positive correlations were revealed between GHDT scores and socioeconomic status in studies conducted in Nigeria (Bakare, 1972) and Turkey ( Uçman, 1972). In contrast, significant differences ware not o b s a ^ between the DAM scores of American Indian children and non-native American children (Cundick, 1970). Similarly, Naglieri ( 1988) found no significant differences in DAP: QSS scores between 250 black-white pairs matched on the basis of collection site, sex, and age. Nor did he find sigiifiœnt DAP: QSS differences between 74 Hispanic-non Hispanic pairs also matched on the basis of collection site, sex, and age. Studies evaluating the effects of examiner and rater differences " n HFD performmce as well those investigating the impact of varying levels of artistic ability among subjects hcve been largely directed toward the projective use of HFDs. These are explored later. The available data on such varioles as they apply to HFD measures designed to assess intellectual development have been reviewed by Anastasi ( 1988). She indicated the effect of the examiner as well as the effect of artistic (Aility were both found to be negligible. The last area reviewed before shifting focus to projective use of HFDs is the impact that psychological adjustment or personality factors has had on HFD tests which were designed to measure intellectual development Ear V findings regarding this impact gave rise to the use of HFDs as a projective measure and, thus, are intimately related to the present study. Goodenough ( 1926) was among the first to suggest emotional states or personality variables influence drawing test performance. In her original study she observed qualitative differences in a small propœtion of drawings which could not be readily accounted for ( pp. 2 0 62-63). She rated nine children whose drawings showed such features with a 5 0 -item adjective checklist and concluded these differences me/ have reflected personality disturbances. Similar reasoning was used to account for lower than expected correlation coefficients in particular HFD concurrent validity studies. For example, Pechoux, Kohler, and Oirard, ( 1947) found a low to moderate association between the DAM and the Stanford-Binet for 8 sample of abnormal and delinquent children. Because this association was lower than that found in previous DAM concurrent validity studies which also used the Stanford-Binet as the criterion variable (i.e., McElwee, 1932; Williams, 1935; Yepsen, 1929), the authors suggested it was a function of the subjects being delinquent However, their sample included children and adolescents up to the age of 18. Thus, both the age factor æd delinquency of the children may have contributed to the lower than expected correlations. Hanvik ( 1953) found a correlation of only. 18 between DAM and WISC IQs for five to twelve year old psychiatric patients with the DAM IQ being consistently lower. Age was obviously less of a factor in this stud/ as this was roughly the age-range deemed most appropriate for the technique when it is used as a measure of intelligence (Sattler, 1988, p. 313). Mortensen ( 1984, p. 58) described Hanvik’s findings as "in line with other results of children with psychic problems." Similarly, Pikulski ( 1972) compared DAM and GHDT scores to WISC scores for 50 boys who were non-achievers in reading, aged seven to eleven. While both DAM and GHDT showed a moderate level of association with the WISC ( .44 and .32 respectively), the A'awing test scores were consistently lower than the WISC scores by about 10 points. Pikulski offered a twofold explanation for this observed difference First, some cf the children appeared to have difficulties with perceptual-motor coordination which influenced botti their drawing and reading performance Second, many of the children were described as anxious and impulsive which made it difficult for them to function in m unstructured task like a drawing test Schaefer and Sternfield (1971) also reported a low to moderate relationship (correlations 21 from .20 to .35) between the GHDT and WISC when testing eight to fourteen year old emotionally disturbed males. Correlation coefficients in each of these studies were lower than many of those reported when the sample was comprised of normals and the consensus among Investigators was that this difference may have been a function of the subjects' emotional or behavioral disturbance. Investigations designed to specifically compare drawing test scores between normal subjects and those with behavioral or personality aberrations have yielded mixed results. Reichenberg- Hackett ( 1953) reported experimentally induced positive affective states influenced children's drawings toward higher scores on the DAM. Ochs ( 1950) compared two groups of behaviorally-disordered c h il* ^ matched for intelligence. One group received hospital treatment and subsequently showed improved social adjustment while the oth«' group showed no change on this dimension. When both groups were retested the improved social adjustment group's mean DAM score had decreased while that of the unchanged group had increased. The resulting difference was statistically significant These findings help provide a better understanding and a fuller appreciation of the number and complexity of factors which influence HFD performance. Cleerly, each of the factors reviewed above must be carefully considered in the administration and interprétatif of HFDs, both as a clinical instrument and œ a research tool. HFD's as a Projective Personality Test As noted earlier, Goodfojÿî ( 1926) was among the first to st^gest there is a relationship between HFD performance and psychological adjustment and she viewed this apparent relationship as worthy of further research. While formal research on the topic did not immediately follow, continued theorizing about it did. Even thouÿi Luquet's (1913) early work focused on aspects of drawing which were related to intellectual maturity, he later 2 2 pointed to changes in children's drawings whidi reflected an emotional realism rather than an intellectual realism (Luquet, 1927,1929). Lewis (1928) (cited in Mortensen, 1984) viewed drawings as more valuable than dreams and attempted to provide systematic descriptions of personality based on drawings. When studying the crayon drawings of elementary school children, Lembke ( 1930) noted observing marked differences between the Sawings of bold and shy children. Much like Luquet, Lowenfeld ( 1947,1952) proposed that the drawirgs of young children reflected the drawer's desires, feelings, beliefs, and fancies rather than objective reality. The most well-known efforts at developing a systematic approach to the interpretetion of HFDs as a projective technique came in the form of the Itouse-Tree-Person Test (H-T-P)(Buck, 1948) and Machover's (1949) Draw A Person Test (DAP). Much like Goodenough's work set the stage for future developments in the use of HFDs as a measure of intellectual maturity, Machover's test served as a foundation for the proliferation of literature focusing on HFDs as a measure of the projected self. Many texts followed which focused on applying these principles (ag,. Hammer, 1958; Koppitz, 1968; and Urban, 1963). vet Machover's Personality Prtnection in the Drawing of the Human Fioure has remained bv far the most quoted source ( Mortensen. 1984). The development of Machover's technique was couched in the study of psychopathology and her own clinics! work with patients. Thus, her hypoUwses grew from her exparience with individuals who were experiencing psychological distress of some nature. Particular focus was on the "strictly personal and individual aspects of the drawings" ( Mortensen, 1984, p. 71 ). This starting point is in sharp contrast to that of the HFD tests designed to assess intellectual maturity, as described earlier. Other important contrasts between these different HFD applications (as a measure of intellectual maturity v k -s u s a projective persorelity technic^) are found in the nature of the actual scoring criteria and the manner in which the results o-e interpreted. For example, Goodenough's ( 1926) DAM is comprised of a group of 2 3 items, each of which has carefully defined scoring criteria. The number of DAK items present constitutes a total score which serves as an estimate of general level of intellectual ability based on its comparison to a normative sample. Machover's ( 1949) DAP consists of general guidelines to identify particular HFD features (or items) which she suggested are associated with specific intrapersonal or interpersonal conflicts. That is, a one-to-one relationship was hyfxithesized to exist between particular HFD signs and specified characteristics or areas of conflict the drawer may be experiencing. Individual signs and the various configurations of these signs were the essential focus of her technique and interpretation was based on a comparison b) theoretical constructs rather than normative data. While the projective use of HFDs has largely conformed to Machover's ideas, one exception is noteworthy. Koppitz' ( 1968), in developing items for her system of HFD Emotional indicators, provided detailed descriptions for each item and used actuarial methods as an attempt to differentiate meaningful from non-meaningful items. Also, in Koppitz' system, the total number of items present per HFD was viewed as interpretively significant rather than the sole focus being on individual items. Generally speaking, individual HFD projective items have been categorized as one of two types: structural/formal, or content The following aspects have been termed structural or formal varioles: overall size, head size, placement on page, perspective, stance, detailing, line quality, shading, erasure, omission, transparency, distortion, symmetry, and action or movement. Content variables are comprised of: face, facial expression, sex of firs t-A ^ n figure, and the presence of lips, mouth, teeth, nose, eyes, eyebrows, ears, hair, contact features (arms, hands, fingers, legs, feet), trunk, breasts, shoulders, hips and buttocks, waistline, joints, props or special themes, anatomy indicators, ambiguously saxed figures, and nudity/clothing. In addition to individual sig e, global or overall HFD characteristics have been recognized as being interpretively meaningful. Global ratings have taken essentially two 2 4 forms. The first consists of a molar or overall impression of the drawing while the second is comprised of a thorough checklist of individual items or signs ( both structural/formal and content) which is viewed in a holistic manner. The empirical findings regarding the technical qualities of HFDs when applied as a projective technique are conflicting and often negative. Although Machover ( 1949) stated her method was valittetsd through innumer^le clinical exsmications and through blind snslysis, she presented no date to support her statement Comprehensive reviews of the HFD literature have been periodically conducted and the technique's reliability and validity have been seriously challenged. The most noted reviews Include the following. Swensen (1957,1968), in separate articles, reviewed the research from 1945 to 1956, and from 1957 to 1966. Roback ( 1968) evaluated the findings over an 18 year period ( 1949 to 1967), and Kahili ( 1984) evaluated the HFD findings for the period from 1967 to 1982. Results of these reviews as well as other relevant findings are presented below. Reliability. Exhaustive reviews covering the period from 1945 to 1982 (i.e.. Kahili, 1984; Swensen, 1957,1968; Roback, 1968) failed to identify even a single study which focused on intra-rater reliability for HFD projective items. However, one sW y was located which evaluated intra-rater item-agreement for Koppitz" ( 1968) Emotiwal Indicators. When a single rater scored and re-scored 60 drawings for Koppitz" 30 Emotional Indicators, with a one-week interval between scorings, 88 percent of the items were scored the same both times (Dieffenbach, 1977). Twenty-eight studies examining inter-rater reliability were reviewed cfcjring the same period and these yielded quite variable results. Swensen ( 1968) reported inter-rato' reliability coefficients for global aspects of HFDs tended to be higher (ranging from .59 to .97) than those for structural aspects (ranging from .23 to .52), and coefficients for content variables were lowest of all. Kahili ( 1984), based upon her review covering a subsequent time period, suggested "respectable"" (p. 271) inter-rater reliability 2 5 coefficients were demonstrated for the majority of HFD projective items. Although there were a few low coefficients the majority were over .80 and three-quarters were .70 or greater. This held true for global, structural, and content variables. Roback did not differentiate among the various categories of HFD items when addressing reliability in his review, but he generally concluded there was a ‘high rate of error made in rating drawings (p. 16).‘ The results of three studies on Koppitz" Emotional Indicators suggested that system possesses a hiÿi of level inter-rater item-agreement Koppitz reported she and another rater agreed 95 percent of the time when independently scoring 25 drawings. The results of similar investigations on the Koppitz system showed inter-rater item-ayeement levels of 81 percent based on the scoring and re-scoring of 60 HFDs (Dieffenbach, 1977), and 91 percent based cn the scoring and re-scoring of 264 HFDs (Klepsch, 1979, cited in Klepsch & Logie, 1982). The available data on rater reliability has been mixed, with some suggestion that global ratings tend to be more reliable than individual signs (structural/formal and content). The apparent inconsistency regarding rater reliability of individual signs may have been a function of different methodological conditions being used Airing the period of early versus more recent reviews. Recommendations based on early reviews of the literature (Swensen, 1957; Roback, 1968) called for the development of more objective scoring procedures. The greater detail evident in Koppitz" ( 1968) criteria for scoring Emotional I ndicators, as compared to the more general scoring guidelines used previously, serves as an example of such development Investigations by Cassel, Johnson, and Burns ( 1958) and Strumpfer ( 1963) empirically demonstrated that explicit scoring instructions and/or scorer training significantly improved the consistency between raters. Thus, the differences observed among studies evaluating the rater reliability of individual sigis may partly reflect a greater effort by investigators to objectify and standardize rating procedures and to adequately train j udges during the period of more recent studies. 2 6 The results of test-retest reliability studies on the HFD technique have also been characterized by inconsistency. Much like his findings on rater-reiiabiiity, Swensen ( 1968) reported global aspects of HFDs (.74 to .90) tended to be more stable than structural aspects (.21 to .85), and content variables showed the least stability of all. KahilTs ( 1984) review examined the results of only four test-retest studies, but reported more consistent reliability coefficients for each type of HFD variable. Tœt-retest reliability coefficients ranged from .81 to .99 across global, structural, and content variables. As with rater-reliability this difference may have been a reflection of the use of more objective rating procedures and better training of scorers during the period of Kahili's review. However, basic principles related to the nature of the HFD technique and the theory on which it was based may also help explain the observed variability in test-retest reliability studies. Koppitz noted that HFDs must not be regarded as a portrait of the drawer's basic and enduring personality traits, but instead as a reflection of the drawer's current stags of mental development and his or her attitudes and concerns of the given moment Because these factors diange over time due to maturation and experience, HFD scores will reflect these dmnges. Leland ( 1983) elabor:;tcd s similar point, noting that should the characteristics of a subject's HFD "... remain st^ le despite environmental modification, this in itself is a sign of noninteraction between client and environment" ( p. 155). Additionally, Mortensen ( 1984) a s s e r t that on one occasion an individual may express a conflict Uirouÿi tl% shading of a body po't ( theoretically denoting anxiety), but later express the same or a related conflict through exaggerating the size or detail of the same body part (theoretically, an attempt at compwsation). Thus, as subtle changes occur in one's environment and in one's style of coping, HFD characteristics change accordingly. In this context, one may conclude that global aspects of drawings may show greater stability than individual signs because global ratings focus on the overall number aid configuration of of signs or items, without regard to their 2 7 particular hypothesized meaning. For example, an attempt at compensation, as well as the presence of anxiety, would contribute to an Increased global score. Alterrate forms reliability has scarcely been evaluated for HFD projective items and the major reviews have not directly addressed this issue. However, two particular studies did offer reliability coefficients which were actually an alternate forms comparison even though that term was not used. Apfeldorf, !tando1f, and Whitman ( 1963) (cited in Roback, 1968) reported coefficients in the .80 s for three particular structural aspects of HFDs. Hammer and Kaplan ( 1964 a, 1964 b, 1964 c, 1966) (cited in Swensen, 1968) found highly conflicting results as to the appearance of various structural and content items across first and second figures drawn by children. Thus, no clear conclusion has been (frawn as to the consistency of perftrmance across individual drawings. However, it should be noted that differential treetment of male and female drawings has been suggested as an impartant item for assessing discrepant attitudes toward men and women. No studies were located which included an evaluation of the internal consistency of HFD projective items, leaving this questim of technical quality unanswered. Evaluation of the evidence regarding the reliability of HFD projective items was complicated by a number of factors ( Kahili, 1984 ; Robadc, 1968). For example, investigators have applied different statistical procedures (ag., reporting percentage of item- agreement versus reliability coeffLients) that were not directly comparable. Additionally, a general lack of standardized scoring procedures failed to make cles^ whether items named similw'V by different authors were measuring the same factor or even what was meant by a particular undefined term. A similar lack in consistency with which raters were trained across studies further complicated the interpretation of results. Despite these complications, the findings and recommendations related to the global rating of (Sawings merit elaboration as they have direct implications for Ü» design of future research. 2 8 Swensen ( 1968) observed that the "reliability of a particular sign is a direct, linear function of the amount of drawing behavior included to assess that sign" ( p. 40). Because global ratings involve all of the drawing behavior in a given HFD, they have tended to be the most reliable aspect of the technic^. Swensen also viewed these global aspects as demonstrating levels of reliability which were considered satisfactory for most psychometric purposes. He further suggested that, when administering the technique, more than one (Rawing should be obtained per subject to provide a larger sample of drawing behavior to evaluate. It follows that the quality of future HFD research can be enhanced by providing more explicitly defined scoring criteria, by offering better detailed descriptions of rater training procedures, and by clearly describing the statistical methods used for evaluation. Additionally, a global rating procedure may provide a more reliable measure of HFD performance as compared to the more narrowly focused consideration of individual siops. A discussion regarding the validity of HFD projective items is now in order, however, a particular issue must first be elucidated. Walsh & Betz ( 1985) emphasized reliability is a "necessary prerequisite" ( p. 55) for validity. "If we lack confidence in the accuracy, consistency, or repeatability of our measurements, there is no point in attempting to relate the instrument to other behaviors, characteristics, or measures" ( p. 55). The foregoing discussion cn the reliability of HFD projective items revealed data which is conflicting at best Thus, it appears investigators evaluating the validity cf this technique have often igna*ed ths fundamental importance of reliability. Such an oversight has blemished the practical significance of such research and, subsequently, failed to establish the foundation necesssy for a technique which receives wide-spread use. With this point in mind, the findings regarding construct and concurrent validity follow. 2 9 Validity. Evalieting the construct validity of HFD technique tes been hampered by the lack of an adequate theoretical construct to relate it to. The ill-defined body-image hypothesis is the construct most frequently considered in this regard This construct was put forth tiy Machover ( 1949) who stated: The human fiffjre drawn Ijy an individual who is directed to "draw a person" relates intimately to the impulses, anxieties, and compensations characteristic of that individual. In some sense, the figure drawn is the person, and the paper corresponds to the environment This may be a crude formulation, but serves well as a wcrking hypothesis, (p. 35) Hammer ( 1959) cautioned against a narrow definition of the body-image hypotlœsis, noting that the figure drawn may reflect one's actual self, one's Idealized self, one's fea'ed self, and/or one's perception of significant others. Investigations designed to evaluate HFD performance in relationship to this construct have yielded mixed results. Swensen ( 1957) identified very few studies prior to 1957 which adequately evaluated the body- image hypothesis. Based on the results of five studies he concluded "definitive research on the basic meaning or sigilficaoce of human fi^ re (Sawings is lacking" (p. 437). Subsequent reviews cmducted by Swenson ( 1968) and Roback ( 1968) examined the results of 15 different studies (eleven reviewed by Swensen, eight by Roback, with four being common to both) which addressed the body-image hyfXJthesis. And, Kahili ( 1984) later reviewed 16 more recent studies on the topic. Methodological ^proaches varied greatly. HFDs were compæed to various measures which were assumed to reflect one's body-image including self-concept scales ( largely of the self-report, sonantic differmtial type), measures of creative expression, and galvanic skin response body gradients (ag,, Armstrong&Hauck, 1961;Flsfier, 1959; Hunt &Feldmæ, 1960;Kanario, 1960; Ludwig, 30 1969; Rosenthal&Beutel, 1981; Van Dyne&Carskatlon, 1978). Blind matching by raters of subjects' HFDs and their full-length photographs was attempted (eg., Apfeldorf & Smith, 1966; Apfeldorf, Smith, Peixotto, & Hunley, 1974). The HFDs of subjects who were physically different, disabled, or disfigured (e g., orthopedically handicapped patients, dermatologie patients, paraplegic individuals, polio victims, obese subjects, and pregnant women) were compared to those of normal subjects (e.g., Bailey, Shinedling, & Payne, 1970; Johnson, 1972; Leichtman, Burnett, &Robinson,1981 ; Manganyi, 1972;Yiney,Aitken,& Floyd, 1974). Also, subject's actual physical height and weiÿit characteristics were compared to the relevant dimensions on their HFDs (ag., Silverstein and Robinson, 1961). Based on his interpretation of the available data Roback ( 1968) concluded; "Although there appears to be some support for Machover's ( 1949) hypotheses, the inconsistent findings indicate that the relationship between figure drawings and body-image is still unclear" ( p. 3). Swensen ( 1968) reached a similar conclusion, having stated there is evidencs that “scores on various aspects of the DAP are significantly related to some other measures that would be expected to reflect a subject's image of himself" ( p. 25). Kahili ( 1984) described the empirical evidence regarding the body-image hypothesis as "... mixed, with slightly more findings failing to support the hypothesis than supporting it" (p. 271 ). During the period of Kahili's review, only seven of sixteen studies supported the hypothesis. Thus, the construct in question has proven difficult to precisely define and attempts made tow»*d demonstrating it's relationship to HFD performmce have been very inconsistent axl, thereby, inconclusive. The importance of whether or not body-image is reflected in HFDs has been challenged. Swensen ( 1968) suggested "... it could be said that a definitive test of the hypothesis is impossible" ( p. 22). Ebei ( ! 961 ) indicated it is a question that should not even be asked. This position argues that whether or not HFDs can be shown to reflect the drawer's 31 true boV"image, however that term may be defined, is virtually meaningless since the HFD’s value lies in its usefulness in predicting certain kinds of behavior in certain kinds of situations (Swensen, 1968). Thus, the question of criterion-related validity is viewed as paramount in evaluating the technique’s usefulness as the construct of body-image may be too diffuse to adequately evaluate and, in the end, does not lend meaningfully to evidence of the technique's technical quality. Before addressing the matter of criterion-related validity, other evidence regarding the constructs underlying HFD performance should be considered. Two factor-analytic studies (Adler, 1970; Nichols & Strumpfer, 1962) were conducted as attempts to identify what factors best explain HFD results. Each study found a single factor which accounted for most of the va'ifflice in figure drawing performance. Nichols and Strumpfer termed this factor "overall quality" (p. 160) while Adler referred to it as the "formal accuracy" ( p. 55) of the drawing. Adler pointed out that the drawing aspects which were reflected in this factor were precisely the same which developed in children with increase in age a d cognitive maturity. Based on this he concluded the factor was primarily one of cognitive maturity or sophistication. Each factor-analytic study also revealed three additional factors which were of much less power. Theauthorsof the two studies did not wholly agree on what these lesser factors represented, but Adler suggested they were “relatively independent of each other, if not completely independent of cognitive maturity" ( p. 57). He also noted they may bear some relationship to significant personality or diagnostic variables and he believed they were deserving of further investigation. Criterion-related validity of the HFD technique has been investigated by examining the association between various HFD items or variables and their hypothesized interpretive siyiificwwe according to Machover ( 1949). The particular HFD aspect examined as well as the type of criterion measure employed have varied yeatly. HFD results in the form of single 3 2 signs (structural and content), overall general Impressions, and formalized global ratings have been compared to the results of other assessment techniques Including Interviews and other formal psychological tests. Attempts have been made to sort the HFDs of disturbed versus normal subjects Into correct categories. Attempts have also been made to correctly sort the HFDs of disturbed subjects into categories differentiating between one diagnostic youp versus another. These methodological approaches were very similar and, in some cases. Identical to those employed to evaluate the HFD's construct validity. This has tended to blur the distinction between these theoretically separate components of test quality and It further reflects the difficulty In precisely Identifying and quantifying the construct being measured. The remainder of this section reviews the empirical findings which are regarded in the literature as Investigations of concurrent validity. Each category of HFD Item or aspect (structural or formal, content, and global ratings) is individually addressed. Machover's specific hypotheses regarding what each individual item or sign represents are not elaborated upon here as they do not pertain to the present study, but the significance she attributed to each sign may be found in her original text or in the articles by Kahili ( 1984), Roback ( 1968), or Swensen ( 1957, 1968). Studies which examined the validity of structural or formal aspects of HFDs have provided conflicting and largely negative results. Swensen (1957, 1968) reviewed the available data on 19 structural items and 17 of these were either not supported by the evidence or the evidence was Inconclusive. Only omission md distortion were found to differentiate between normals aid severely disturbed subjects. The omission item was defined by the omission of significant body parts aid, while effectively differentiating betweai normals and severely regressed subjects, it did not distinguish between normals and near-normals (Baldwin, 1964; Handler & Reyher, 1964,H11er&Nesvig, 1965;Krfin& Jores, 1965; Koppitz, 1966a, 1966b, 1966c; Lorge, Tudcman, & Dunn, 1958; Mogo*, 3 3 1962; Rosenberg, 1965; Vane&Eisen, 1962). The distortion item was defined as bod/parts drawn out of proportion, parts not connected to the bod/, and parts drawn in inappropriate areas of the bod/. It was judged as the most convincing sign of maladjustment among structural items in a number of studies ( htendler & Reyher, 1964; Hiler & Nesvig, 1965; K^n& Jones, 1965; Koppitz, 1966a, 1966b; Lorge, Tuckerman, & Dunn, 1958; Vane & Eisen, 1962). Despite this evidence, omission and distortion were among the seven items not sup[»rted in Kahili's ( 1984) review of subsequent research. Kahili additionally observed conflicting findings for the remaining six structural or formal items she examined. Roback ( 1968) reported inconsistent and non-supportive findings in his review of the data on all six of these items he reviewed. Thus, the preponderance of evidence has failed to support structural or formal HFD signs as valid indicators of behavioral characteristics associated with emotional disturbance. Table One summarizes the results on specific HFD structural and formal items. Swensen ( 1957,1968) and Roback ( 1968) described the results of studies evaluating the validity of more than 30 content variables as conflicting and negative. None of the items were supported by the availdjle d«?ta with the exception of the sex of the first-drawn figure. This it«n was the focus of 18 studies during the 1957-1966 period while no other content item was the subject of more than six studies. Many content variables were examined in as few as one or two studies. Mortensen ( 1984) noted that it is difficult to comf(rt^iy draw definite conclusions based on such limited data. However, Kahili ( 1984) also found the evidence conflicting and failing to support the validity regarding 15 of the 16 content items she reviewed, including the sac of the first-drawn figure. A single study ( Janzen & Coe, 1975) reported significant results suggesting the shape of the trunk effectively differentiated between homosaximl and heterosexual women. Much like the structural HFD items, an 3 4 Table 1 Empirical Findings ftr Validity of HFD Structural or Formal Items Findings Item Swensen* Swensenb Roback Kahili 1) Size + / - + /- ♦ / - */- 2) Head Size * + /- * + / - 3) Head Length « « 4) Head/Bod/Ratio + / - « « 5) Placement ♦ /- 6) StMice + / - + / - « 7) Profile ♦ / - ♦ /- * 8) Detailing « ♦ /- « ♦ / - 9) Line Quality + / - + / - + /- + / - 10) Shading + / - + / - 11) Erasure + / - 12) Omission ♦ « 13) Distortion + * 14) Symmetry « + / - 15) Transparency ♦ / - « 16) Action/Movement + / - « « « 17) Midline Emphasis * * * 18) Unusual Succession * * « 19) Differential M/F Tx * + / - * * : Greater nwnter of studies supporting than not supporting «(1957) - ; Greater nunAer of studies not swporting than swporhng b(1968) */- : Inconciusive findings due to mixed or conflicting results * : No studies reported during this period or item not reviewed 3 5 overwhelming majority of the evidence refuted or failed to support content items as valid reflections of emotiorail dislurtjaœe. Table Two summarizes the results on specific HFD content items. Table 2 Emoirlcal Findings for Validity of HFD Content Items Findings Item Swensen» Swensenb Roback Kahili 1) Head - + / - - + / - » * - 2) Face - 3) Facial Expressloi + / - * - + / - 4) Eyes + / - - ♦ / - */- 5) Eyebrows/Lashes * - - - 6) Nose - - - + / - 7) Mouth ♦ / - + / - * + / - 8) Lips + / - « - + / - 9) Teeth « + « + / - 10) E«^ - - + / - - 11) Hair ♦ / - ♦ + - 12) Neck « - ■if + / - 13) Shoulders « + * - 14) Arms ♦ / - */- - ♦ / - 15) itends ♦ / - - - ♦ / - 16) Fingers * + / - - + / - 17) Lsp - - - + / - 18) Joints - + + ♦ / - IS) Feet --- + / - 3 6 Table 2 (continued) Findings Item Swensen® Swensenb Roback Kahili 20) Toes « ♦ « + / - 21) Limb Size * - « ♦ / - 22) Trunk * + / - * + 23) Breasts - * ♦ / - + / - 24) Waistline ♦ / - * - + / - 25) Hips/Buttocks « - - ♦ / - 26) Sex Drawn First « ♦ « - 27) Sex Symbols - - * - 28) Sex Ambiguity »« « ♦ / - 29) Nudity * ♦ / - * + / - 30) Clothing «« ♦ / - + / - 31) Props • « « - 32) Earrings « - « ♦ / - 33) Buttons * - « ♦ / - 34) Belt « - « ♦ / - 35) Heels « - « + / - + : Greater number of studies supporting thai not sifporting «(1957 ) - : Greater nunter of studies not supporting than supporting *>(1968 ) *!- : inconclusive findings due to mixed or conflicting results * : No studies reported during this period or item not reviewed 3 7 Olobsl ratings have been the most encouraging aspect of HFDs regarding their application as a projective technique even though the eviffence here, too, was mixed Swensen ( 1957,1968} indicated that global ratings "do not significantly detect specific kinds of pathology," but do significantly relate to variables whi(* are "reflections of gross maladjustment" (p. 28). Swensen ( 1968) argued that global ratings of HFDs tend to be more valid in comparison to structural or content items because of the "relative unreliability" (p. 28) of the structural or content items. Kahili ( 1984) described the evidence on global ratings as "mixed" (p. 285), but noted that a number of global measures have been shown to successfully differentiate between normal and pathological groups. She additionally reported that global measures have displayed significant relationships to various criterion measures among ntrmal aid clinical populations. A sampling of the pertinent evidence follows. Global HFD characteristics were shown to correlate significantly with various scales on the hinneso(B riultiphasic Personality Inventory (MMPI) (Hathaway & McKinley, 1943) when studying 89 psychiatric hospital patients. The sample was comprised of 42 males and 47 females with various diagnoses and correlations ranged from .25 to .32 between the global HFD ratings and the MMPI F scale, L scale. Ego Strength scale. Sexual Deviation scale, and Psychopathic Deviancy scale (Lewinsohn, 1965). Diagnoses based on the H-T-P were also shown to be significantly related to MMPI diayioses (Pryor & Butlar, 1969). Investigations byJegedeandBamboye( 1981)and0ttenbacher ( 1981) demonstrated that overall HFD ratings correlated significantly with self-concept scales. Russakoff, Fontona, Dowds, and Harris ( 1976) found a significant association between global HFD scores and subjects' ratings on psychological adjustment as judged by psychotherapists. The HFDs of 104 subjects who applied for admission to a psychiatric hospital were rated for global HFD characteristics. Only 60 of the ^ lic a n ts were actually admitted to the hospital. The global HFD scores successfully predicted (fi < .0001 ) which applicants would be subsequently admitted ( Kahn & 3 8 Jones, 1965). Koppitz ( 1968) compared the HFDs drawn by a group of children from a child guidance clinic to those drawn by a group of well-adjusted children and found an overweight of Emotional Indicators in the drawings produced by the clinic patients. Judges who attempted to sort the HFDs of adolescents into normal and psychiatric patient groups based on global ratings correctly did so 66 to 73 percent of the time (Strieker, 1967). Numerous other investigations have offered similar support for the validity of global ratings as a reflection of emotional disturbance (e.g., Carlson, Quinlan, Tucker, & Harrow, 1973; Cauthen, Sandman, Kilpatrick, & Deabler, 1969; Handler&Rayher, 1966; Haworth & Normington, 1961; Lair & Trapp, 1960; Lapkin, Hillaby, & Silverman, 1968; Maloney & Olasser, 1982; McLachlan &Head, 1974; Nichols & Strumpfer, 1962; Phillips & Phillips, 1976; Rabin & Limuaco, 1959;Silverstein, 1966). The practical significwce or clinical utility of some HFD validity studies which have demonstrated statistically significant results has been challenged. For example, while Koppitz' ( 1968) study found an overweight of Emotional Indicators in the clinic patients' drawings, a few clinic patients showed the presence of no Emotional Indicators on their HFDs and the HFDs of a few well-adjusted subjects showed a number of Emotional Indicators similar to that of the majority of clinic patients. Thus, the absence of these Indicators did not rule-out emotional problems, nor did the presence of a few Indicators necessarily denote disturbmce. The results of Strieker's ( 1967) study were questioned by the author himself. While judges correctly classified the HFDs of normal versus psychiatric patients with 66 to 73 percent accuracy, the rate of correct judgments did not vary meaningfully from the base rate of 57 percffit had all the drawings been classified as pathological. Strieker concluded that, althouÿi statistically sigiificait, the results were of little practical value. Kahili ( 1984) indicated that most classification attempts like that of Strieker have either failed (e.g., Rubin, Ragins, Schachter, & Wimberly, 1979; Wanderer, 1969; Wildman & WilAnan, 1969), or provided a 3 9 degree of improvement tarely above chance and, therefore, meaningless (Cressen, 1975; Strieker, 1967; Watson 1967). The results of other studies have plainly failed to support the validity of global HFD ratings. Global HFD clwacteristics failed to show a significant relationship with social conformity (Sol»', Bruehl, and Kovacs, 1970) aid with neurotic and psychotic profiles on the Sixteen Factor Personality Questionnaire ( I6PF) (Cattell, Eber, & Tatsucka, 1970) (Marzolf&Kirchnar. 1972). Inan investigation designed to evaluate the association between HFDs and several assorted criterion measures, global characteristics were not significantly related to Rorschach ( 1942) variables, Rotter's ( 1966) locus of control, or clinically judged depersonalization (Carlson, Tucker, Harrow, & Quinlan, 1971). Evaluation of the evidence regarding the validity of HFD projective items has been complicated by maiy factors similar to those which affected the evaluatioi of reliability data. These include investigators applying different methodological approadies that were not directly comparable ( e g., attempts to relate HFDs to very different criterion measures), a general lack of uniform administration md scoring procedures, and a similar lack in consistKicy with which raters WK'e trained across studies. The fact that many aspects of the HFD could not te relitely scored further complicated the validity issue. Howevo', as was true with the reliability data, particular findings and recommendations regarding the validity of global rating ma'it further consideration as they have clear implications for future research. Amidst conflicting empirical findings, the global rating of HFDs has been judged by maiy as the most promising aspect of the technique. Those who have conducted extensive reviews of the literature (i.e.. Kahili, 1984;Swenæn, 1957,1968; Roback, 1968), as well as those who have evaluated narrower portions of it ( i.e. Handler & Re/her, 1965; Klopfer & Taulbee, 1976; Suinn & Oskanp, 1969), essentially agreed that global ratings speared to be more 4 0 valid as an indicator of emotional disturbance than either structural or content variables. The majority of evidence suggested that, while these ratings did not successfully distinguish normal from near-normal subjects, they did significantly differentiate between gross levels of adjustment versus maladjustment Despite this evidence, the practical value of the technique appears to be another matter. While statistically significant results were often reported regarding the effectiveness of global ratings In differentiating youps of normal and maladjusted subjects, the ^ illty to correctly predict group membership of an Individual subject based on these ratings sometimes failed to differ meaningfully from chance. Several considerations for the improvement of future HFD research seem to grow from these findings. First, because global ratings appear to be a more reliable and a potentially more useful index of behavioral or emotional disturbance than Individual Items, such ratings should be the primary focus of HFD research. Second, as HFDs seem to have been more successful at differentiating between gross levels of adjustment versus maladjustment, the technique should be used as a screening measure to assess these types of differences rather than to make fine discriminations between, for example, ncrmals and near-normals. Third, criterion measures ^wuld be carefully thought out and described Last, when significant group differences are found, the rate with which individual subjects can be correctly classified should be explored Confounding variables. Several factors extraieous to psychological adjustment have been investigated In altanpt to further Identify variables which Influence a subject’s performance on these HFD measures. It has been clearly shown that intellectual maturity accounts for a 1 « ^ portion of the variance In these drawings. This area was reviewed earlier and will not be repeated Other factors examined as to the impact they have had on HFD performance Include a number of subject characteristics ( I.e., age, sex, race, artistic ability) as well as several situational varioles ( i.e., administration conditions, Individual judge or 41 rater characteristics). The available evidence evaluating the Impact of the variables most frequently considered In this regard Is reviewed below. The mejorlty of research on sex differences has focused on a single HFD aspect, this being the sex of the first-drawn figure. Heinrich end Triebe ( 1972) found, after reviewing 19 studies on the topic, that both boys and girls tended to draw their own sex first when given tt« option. This trend was true for all ages and became stronger as children matured ( Blellauskas, 1960, Tolor & Tolor, 1974). Studies by Hammer and Kaplan ( 1964 a) and Litt and Margoshes ( 1966) Investigated the reliability of first-drawn sex. Both Investigations found that subjects who drew opposlte-sex figures first on one occasion did not necessarily do so upon subsequent HFD adminlstratlon(s). These authors urged caution In the Interpretation of the Item first-drawn sex. Other miscellaneous findings on sex differences Include that boys produced more profile drawings than girls ( Hammer & Kaplan, 1964b; Morris, 1955), more girls tended to omit or hide essential body parts ( e g., «'ms and hands) ( Morris, 1955), and girls tended to more frequently draw suns on their HFDs ( Loney, 1971). A few specific HFD aspects have been ^wwn to v a y with the age of the drawer. As subjects increased in age they spent less time on their drawings and, when given the option of making chromatic HFDs, their use of bright colors also declined with age ( Morris, 1955). Zuk ( 1962) found that the helÿit and width of A'awlngs tarded Increase with age, and Loney (1971) reported finding suns drawn more frequently on the HFDs of young children. Findings on drawing differences due to race have been conflicting. Hanmer ( 1953) compared the H-T-P drawings of normal black and white children and reported finding more characteristics of malacÇustment, hostility, and agression among the black subjects. An attempt to replicate Hammer's study was made by Kuhlman and Blellauskas ( 1976). They matched their groups for Intelligence and socioeconomic status and found no significant black/white differences In H-T-P ratings for Intellectual ability or for personality 4 2 adjustment McHugh ( 1963) studied the drawings of black, white, and Puerto Rican children and found that the Puerto Rican children tended to draw smaller figures. However, no differences among the groups were found to reflect discrepant levels of hostility or aggression. The impact of HFD administration conditions has been evaluated by altering these conditions in various ways. Shanan ( 1962) explored the effect of specific versus ambiguous instructions and found greater variability among the (Nwings of subjects who received ambiguous directions. Cassel, Johnson, and Burns ( 1958) evaluated the Impact of having the examiner present versus the examiner being absent and found a greater number of deviant signs on the HFDs of subjects who àr&N with the examiner absait from the room. Goldstein and Rawn ( 1957) found an increase in signs of aggression when subjects (state hospital attendants) were told they would have to work longer hours without an increase in pay. Exner ( 1962) found a that a youp of normals in which fear was irxfcjced by puncturing their fingers for a blood sample drew as much shading in their drawings as a group of neurotic subjects. Handler and Reyher ( 1964) also found a significant increase in the number of anxiety indicators on HFDs produced by subjects’ who were placed in stressful situations. The effect of various judge or rater cha'acteristics has been richly investigated. Numerous studies demonstrated that the level of prior experience of judges in rating HFDs had little to do with the accuracy of their ratings. Maiy comparisons have been made between psychologists and non-psychologists (teachers, school administrators, middle school and university students, secretaries) as to their ability to differentiate between normal and disturbed samples and between different pathological groups. The evidence has consistently shown no significant difference among these group's accuracy of rating ( Arkell, 1976a; Burton &Sjoberg, 1964,-Oessen, 1975; Feldman & Hunt, 1958; Hiler & Nesvig, 1965; Schaeffer, 1964; Strieker, 1967; Ziv & Shechori, 1970). Investigations by Schmidt and McGowai ( 1959) and by Murray and Deabler ( 1958) similarly found that experience did not 4 3 improve psychologist's ability to sort the HFDs of psydiiatric patients into correct diagnostic categories. Hammer and Piotrowsky ( 1953) examined the effect that raters' personality factors had on HFD performance and found that, when scoring HFDs for the presence of hostility, subjects' scores were significantly related to the raters' own level of hostility. Two limitations must be noted regaxiing the majority of these investigations on the effect of rater characteristics. First, judjnents were often based on the rater's general impression of the HFDs rather than a formal scoring system. Second, many of the studies attempted to make fine distinctions between subjects, for example, those diagnosed with undifferentiated schizophrenia versus those diagnosed with paranoid schizophrenia (ag., Murray & Deabler, 1958) rather than focusing on gross levels of adjustmoit versus maladjustment The problems associated with such practices have already been discussed. The influence of subject's overall artistic * ility on HFD performance has been widely studied. The evidence has consistently demonstrated a significant relationship between psychologist's ratings for psychological adjustment æd the ratings of a'tists or art teachers' for artistic quality ( Bieliauskas & Bristow, 1959; Feldman & Hunt, 1958; Marias axl Strumpfer, 1965; Sherman, 1958a, 1958b; Solar, Bruehl, & Kovacs, 1970; Whitmyre, 1953). In evaluating the evidence available at the time of his review, Swensen ( 1968) indicated that these studies only proved that judgment of artistic excellence and judgnent of psychological adjustment are cwrelated, not that these elements are one in the same. He did, iTOwever, recommend that future HFD research control for drawing ability. Summary The Human Figure Drawing technique has been widely investigated both as a measure of intellectual maturity and as a method of projective personality assessment This review of the relevait reli^ility and validity data revealed satisfactory results regarding its use as a 4 4 nonverbal screening measure for cognitive ability. However, the data evaluating the technique's usefulness as an indicator of psychological or emotional adjustment was conflicting and less supportive. Several rœommendations emerged from this past research on the use of HFDs as a projective technique which have implications for future research. These centered around the need for more objective rating procedures, a shift in focus toward the global aspects of HFDs end away from individual sign interpretation, and applying the technique as a screening measure for gross levels of adjustment rather than as an indicator of a specific conflict or anomaly. Additionally, several confounding variables were identified which must be considered during the clinical interpretation of HFDs. These variables should also, when feasible, be controlled for in future HFD research. Chapter III Methodology This study examined the feasibility of using Human Figure Drawings ( HFDs) to screen chilcfren for potential emotional or behavioral disorders. A scoring system, or scale, to rate HFDs procfcjced by children was developed as a portion of the study. The scale was a modification of existing approaches found in the literature wd its construction reflected a combination of rational and empirical methods. The technical qualities of the scale were closely examined: intra-rater, inter-rater reliability were evaluated as were internal cmsistency, concurrent and discriminant validity. Subiects Two youps of subjects were studied. The first group, ta'med the normal group, was comprised of 186 seven to thirteen year old males selected from the sample used to standardize the Draw-A-Person: A Quantitative Scoring System ( Naglieri, 1988) and the Matrix Analogies Test - Short Form (Naglieri, 1985). This standardization sample, collected in the fall of 1984, is nationally representative according to the 1980 United States Census Bureau data on the following variables; age, s®<, ethnic group, geographic region, and community size. Only children enrolled in regular ( not special education) classrooms participated in the standardization process. The normal youp for the current project included whites md blacks selected fron this standardization sample and stratified according to race and geographic region of the United States. Four geographic regions were identified (Northeast, North Central, South, and West), consistent with the 1980 United States Cmsus 45 4 6 Bureau data. The normal group's characteristics are described in Tables Three and Four. Tables Normal Group Characteristics bv Region «TdAoe Geoyaphic Region Northeast North CHitral South West Total Age Yrs/Mths n(% ) n(% ) n(55) n(% ) n(% ) 7 /0 -7 /5 3(2) 4(2) 5(3) 3(2) 15(8) 7/6-7/11 3(2) 4(2) 5(3) 3(2) 15(8) 8 /0 -8 /5 3(2) 3(2) 6(3) 3(2) 15(8) 8/6-8/11 3(2) 4(2) 6(3) 3(2) 16(9) S /0 -9 /5 3(2) 5(3) 6(3) 3(2) 17(9) 9/6-9/11 3(2) 3(2) 5(3) 4(2) 15(8) 10/0-10/5 3(2) 3(2) 5(3) 3(2) 15(8) 10/6-10/11 3(2) 4(2) 5(3) 5 (3) 17(9) 11/0-11/5 4 (2) 4(2) 4(2) 2(1) 14(8) 11/6-11/11 3(2) 4(2) 6(3) 3(2) 16(9) 12/0-12/5 4(2) 3(2) 5(3) 2(1) 14(8) 12/6-12/11 3(2) 4(2) 7(4) 3(2) 17(9) Total 38 (20) 45(24) 66 (35) 37(20) 186 (100) Percentage in U.S. Population* (21) (26) (34) (19) (100) *Based on T^le 55, U.S. Bureau of the Census, 1980 Census of Proulation. PC80-1 -B 1,1983 4 7 Table 4 Normal Qroup Characteristics bv Race and Aqb Race White Black Total Age Yrs/Mths n(% ) n(%) n(% ) 7 /0 -7 /5 12(6) 3(2) 15(8) 7/6-7/11 12(6) 3(2) 15(8) 8 /0 -8 /5 12(6) 3(2) 15(8) 8/6-8/11 14(8) 2(1) 16(9) 9/0-9/5 15(8) 2(1) 17(9) 9/6-9/11 11 (6) 4(2) 15(8) 10/0-10/5 13(7) 2(1) 15(8) 10/6-10/11 14(8) 3(2) 17(9) 11/0-11/5 12(6) 2(1) 14(8) 11/6-11/11 13(7) 3(2) 16(9) 12/0-12/5 13(7) 1 (1) 14(8) 12/6-12/11 14(8) 3(2) 17(9) Total 155(83) 31 (17) 186(100) Percentage in U.S. Population* (79) (15) (6 S -Other') ♦Based on Table 43, U.S. Bureau of the Census, 1980 Census of Populaticffi. PC80-1 -B 1,1983 4 8 The second youp, termed the exceptional group, was comprised of 92 seven to thirteen year old males who were selected on the basis of their having been placed in a special setting (a special classroom in a public school or a residential school) designed for the treatment of emotional and/or behavioral disorders. Each exceptional group subject had been formally diagnosed as having a severe behavior handicap as described in rule 3301-51 -01 of the Administrative Code presented in the Rules for the Education of Handicœœd Children (Ohio Department of Education, 1982). The relevant criteria for this dignosis are presented below. AAA. "SEVERE BEHAVIOR HANDICAP" is defined as follows: 1. The term means a conditiwi exhibiting one or more of the following characteristics over a long period of time and to a marked degree, which adversely affects educational performance: a. An inability to learn, which cannot be explained by intellectual, sensory or health factors; b. An inability to build or maintain satisfactory interpersonal relationships with peers or teachers; c. Inapprt^riate types of behavior or feelings under normal circumstaraes; d. A gena^al pervasive mood of unhappines or depression ; or e. A tendency to develop physical symptoms or fears associated with personal or school pnAlems. 2. The term does not include children who æ e socially maladjusted, unless it is determined they are severe behavior handicapped, (p. 11 ) 4 9 1. Eligibility A child who meets the definition of severe behavior handicapped in paragraph AAA. of rule 3301 -51 -01 of the a*ninistrative code a«l the following requirem&iîs shall be eligible for special education programming and related services fcr severe behavior handic^ped children. a Each child shall have a multifactored evaluation for initial placement that includes, but is not necessarily limited to, evaluations In the following a^eas: ( i) Physical examination completed by a licensed doctor of medicine or doctor of osteopathy; (ii) Vision, hearing, and motor abilities; (111) Communicative status; ( iv) General intelligence as deta'mined throuÿi a mæsure of cognitive functioning administered by a qualified psychologist using a test designed for individual administration; (v) Academic performance; (vi) Background Information inclusive of educational, fanily and medical history; (vii) Informal behavior observation by the child's current teacher and at least one other team member ; and (viii) Bdiavior or personality measure. ( pp. 62-63) Six data colWion sites were used. Five collection sites were located in central Ohio and one was located in western Pennsylvania The investigator administered the appropriate instruments to these subjects in a group format. The exceptional group's characta'istics are described in Tables Five and Six. 5 0 T^leS Exceptional Group Chyacteristlcs bv Oollection Site and Aoe Collection Site 1 2 3 4 5 6 Total Age Yrs/Mths n(% ) n(% ) n(X ) n(%) n(%) n(%) n(%) 7 /0 -7 /5 1 (1) 0 (0) 0(0) 0(0) 0(0) 0(0) 1 (1) 7/6-7/11 3(3) 2(2) 1 (1) 1 (1) 0(0) 0(0) 7(8) 8/0-8/5 0(0) 0(0) 0(0) 0(0) 1 (1) 1 (1) 2(2) 8/6-8/11 0(0) 5(5) 1 (1) 0(0) 0(0) 0(0) 6(7) 9 /0 -9 /5 5(5) 1 (1) 0(1) 1 (1) 1 (1) 0(1) 8(1) 9/6-9/11 2(2) 3(3) 2(2) 1 (1) 0(0) 1 (1) 9(10) 10/0-10/5 4(4) 4(4) 2(2) 1 (1) 0(0) 0(0) 11 (12) 10/6-10/11 9(10) 2(2) 0(0) 1 (1) 0(0) 1 (1) 13(14) 11/0-11/5 2(2) 4(4) 1 (1) 0(0) 0(0) 2(2) 9(10) 11/6-11/11 6(7) 1 (1) 4(4) 1 (1) 0(0) 0(0) 12(13) 12/0-12/5 2(2) 4 (4) 0(0) 2(2) 0(0) 0 (0) 8(9) 12/6-12/11 2(2) 0(0) 1 (1) 2(2) 0(0) 1 (1) 6(7) Total 36 (39) 26(28) 12 (13) 10(11) 2(2) 6(7) 92 (100) 51 Table 6 Exertional Group Ctw acteristics by Race and Age Race White Black Total Age Yrs/Mths n(% ) n (%) n(% ) 7 /0 -7 /5 1 (1) 0(0) 1 (1) 7/6-7/11 4(4) 3(3) 7(8) 8 /0 -8 /5 2(2) 0(0) 2(2) 8/6-8/11 4(4) 2(2) 6(7) 9 /0 -9 /5 7(8) 1 (1) 8(9) 9/6-9/11 6(7) 3(3) 9(10) 10/0-10/5 6(9) 3(3) 11 (12) 10/6-10/11 12(11) 1 (1) 13(14) 11/0-11/5 5(5) 4(4) 9(10) 11/6-11/11 8(9) 4(4) 12(13) 12/0-12/5 6(7) 2(2) 8(9) 12/6-12/11 6(7) 0(0) 6(7) Total 69 (75) 23 (25) 92(100) Instruments Human Figure Drawing-. Futures of Disturbance Scale (HFD: F PS). The development md evaluation of this scale was a primary focus of the current study. These aspects of the scale are detailed in the Procedures and Data Analysis section below. The scale was designed as a screening measure to identify emotional or behavioral disturbaice in children. Use of the 5 2 scale reclines the subject to produce three thawings ( man, woman, and self) under standardized conditions which ^pear in Appendix A. It ma/ be administered individually or to a group and the resultant drawings are scored accordir^ to specific guidelines appearing in Appendix C. The scale is comprised of 31 items which are scored as either present or absent (one or zero) for each indivicfcjal drawing. Thus, each item raw score may vary from zero to three. One total raw score is provided which is the sum of the item scores. This total raw score may vary from zero to 93. Draw A Person: A Quantitative Scoring &/stem ( DAP: OSS). The DAP; OSS (Naglieri, 1988) is a measure of intellectual ability which requires the subject to produce three drawings (man, woman, and self) under staxJa^ized conditions. It may be administered IndlvickJally or to a group and the résultait drawings scored according to specific guidelines spearing in the test manual. The instrument yields standard scores ( Mean = 100; SD = 15) for each of the three individual drawings and for the total test. Reli^ility data reported in the test manual are satisfactory. Internal consistency coefficients raiged from .94 to .96 across tte 13 age gwips (oneyear intervals, ages 5 to 17) with a median coefficient of .95. Preliminoy data on the total test score revealed a median test-retest (four-week interval) reli^iility coefficient of .61 (grades 1-7), inter-rater re lia ility coefficients of .95 (yades 1-3) and .93 (grades 5-7), and intra-rater relW lity coefficients of .97 (grades I-3 ) and .93 (grades 5-7). Validity data reported in the test maiual were also satisfactory. Construct validity was demonstrated by an age-to-age progression of total raw scores, and a factor analysis revealed high item loadings on a general abilities factor. Criterion-related validity was shown throucfi signifient correlations with other standndized measures of ability n d achievement. A correlation coefficient of .37 (g. < .01) was found between the DAP: OSS and the Matrix Analogies Test - Short Form ( MAT-SF) ( Naglieri, 1985) for a group of 594 kindergarten throuÿï third-yade stixjents and of .27 (g 5 3 < 01) for 8 group of 1,328 forth through twelfth-yade students. Similarly, the DAP: OSS was shown to correlate significwitly with both the reading and math sections ( r = .24, £ < .01 and r = .21, a < .01, respectively) of the Multilevel Academic Survey Test ( MAST) ( Howell, Zucker, & Morehead, 1985) for the same group of 1,328 forth through twelfth-grade students. A significant correlation has also been demonstrated between the DAP: OSS and the Goodenough-Harris Drawing Test (Harris, 1963)(r = .87,a< .01) for a group of 100 nine year olds. Matrix Analogies Test-Short Form (MAT-SF). TheMAT-SF (Naglieri, 1985) is a measure of intellectual ability which uses abstract visual analogies of the progressive matrix type. !t too, may be aAninistered individually or to a group. The instrument yields total test scores in the form percentile ræ ks, stanines, and age equivalents. A conversion table ( Sattler, 1988, Table BC Dwas used to allow the expression of MAT-SF results as standa^ scores ( Mean = 100; SD = 15) to facilitate the data analyses for this study. Reliability data reported in the MAT-SF manual were satisfactory. The internal consistency coefficients ranged from .63 to .89 across the studied age groups with a median coefficient of .83. The reliability coefficient for the test-retest (four week interval) sanple examined was .78. Validity data reported in the manual are also satisfactory. Construct validity was demonstrated by an age-to-age progression of total raw scores and a factor analysis revealed high item loadings on a general abilities factor. Criterion-related validity was shown through a significant correlation ( median r = .51, fi < .001 )between the MAT-SF aid the Multilevel Academic Survey Test (MAST) (Howell, Zucker, & Mw'ehead, 1985) and through a si^ificait correlation ( r = .68, < .001 ) between the MAT-SF and the Performance Scale of the Wedisler Intelligence Scale for Children - Revised (WISC-R) (Wechsler, 1974). 5 4 Procecftjres and Data Analysis The procedures for executing this stud/ followed m orderly sequence. Items for the scale were developed by gathering items from the literature and modifying them with the aim of increasing the objectivity with which they could be scored. Raters were trained to use the resulting Prelimirwy Scoring Manual (Appendix 6) and the drawings from the normal group were then scored. The technical properties of each item were examined and the number of items was reduced based upon these properties. The remaining list of itans comprised the Human Figure Drawing; Feetures of Disturbance Scale ( HFD: FDS). Rater-reliability of the scale was examined. The association between the HFD; FDS and two standardized measures of intelligance was investigated for the normal gxnup to determine the divergent validity of the scale. Lastly, the (Sawings from the exceptional group were scored using the HFD; FDS and the scale was evaluated as to its ability to differentiate between groups, a* its concurrent validity. These procedures and the accompanying methods of data analysis are elaborated upon below. item Develroment A list of 94 HFD items was identified to include on the Preliminary Scale. Each item had been suggested in the literature to reflect emotional or behavioral disturbance. That is, each item was crawn from earlier HFD scales or scoring systems designed to detect emotional or behavioral problems. A vaiety of sources were reviewed in forming this collection of preliminary items (Adla, 1970; Daum, 1983; Engles and Suppes, 1970; Hamma, 1958; Hamma and Kaplan, 1966;Handla, 1967; Holmes and Stephens, 1984; Holzberg and W exla, 1950; Hoyt and Baon, 1959; John, 1974; Klepsch and Logie, 1982; Koppitz, 1968;L^in, 1956; Lehna and Gunderson, 1952; Machova, 1949; McHugh, 1966; Montague and Prytula, 1975; Royal, 1949; Saani and Aza^, 1977; and Vane and Eisen, 1962). F a the purpose of this study the items were categaized in one of two classes which were rouÿnly consistât with classes depicted in the literature. The first class, structural items, was comprised of those items whidn could be physically measured. The 5 5 second class, content items, consisted of items which could merely be scored as present or absent. Each item is described in detail in the Preliminary Scoring Manual (Appendix B). Rater training and scoring the normal group (^awinos. Four raters were trained to score the (feta for this project using the Preliminary Scoring Manual appearing in Appendix B. A three-hour group training session was condwted where each item description was reviewed end supervised practice took place. Prior to scoring the actual data collected for this stud/, each rater was required to pass a competency exam to assure the items were being scored accurately. The competency exam was comprised of 10 drawings (four man, three woman, and three self drawings) selected from those which were collected for, but not used in, the actual study. Satisfactory performance was defined as greater than 90 percent agreement between each rater’s scorir^ aid the correct scoring predetermined by the author and his advisor. The raters first scored the HFDs from the normal group. A set of 25 drawings ( nine male, eight female, and eight self-(feawings) was selected to be scored and re-scored by a single rater, with a ten-day period between scoring time-one aid scoring time-two, for subsequent consideration of intra-rater item-agreement and intra-rater reli^ility. Another set of 25 drawings ( nine male, eight female, and eiÿit self-drawings) was scored by two separate raters for subsequent consideration of inter-rater item-agreement and inter-rater reliability. Item analysis and reductioi. The first goal of the item analysis was to identify which items occurred with less than 16 percent frequency in the normal gixjp. This proceAire was conducted in two stages; first for the structural items ( items which could be physically measured) and, secondly, for the content items ( items which could merely be scored as present or absent). 5 6 The drawings were scored for each structural item by carefully measuring the item according to the Preliminary Scoring Manual (Appendix B). These actual measurements were recorded and a frequency distribution was constructed for each structural item for the male, female, md self drawings. Oit-off scores were established for scoring each structural item by identifying the point on each frequency distribution where less thaï 16 percent of the normal group fell. Thus, cut-off scores were created separately for the man, womm, and self (drawings md these structural items were thereafter scored as one (present) or zero (absent) for each drawing, based on whether or not their measurements exceeded the cut-off criteria The drawings were also scored for the presence (one) or absence (zero) of each content item. A frequency distribution was constructed for each content item for the male, female, and self drawings. The mean occurrence rate of each content item, across all three drawings, was calculated aid only content items with a mean occurrence rate of less than 16 percent remained on the Preliminary Scale. The second goal of the item analysis was to further rethjce the numbo' of items based on an evaluation of each item's ability to be rated consistently ( percentage of item-agreement under intra-rater aid inter-rater conditions). Intra-rater item-agreement was based on the percent of item-agreement when a single rater scored the sane drawings on two s e p o ^ occasions, that is, the percentage of items which were scored the sane both times, inter- rater item-agreement was based on the percent of item-agreement between two different raters scoring the same (Sawings, that is, the percentage of items which were scored the sane by both raters. Separate sets of 25 drawings ( nine male, eight female, and eight self-drawings per set) were used to examine the percent of item-agreement under intra-rater and inter-rater conditions. Only items which were Ale to be rated consistently under intra-rater and inter-rater conditions ( > 80 % agreement) remained on the Preliminary Scale. 5 7 The third goal of the item analysis was to further reduce the number of itons based on each item's relationship to the scale as a whole ( item-total correlation). Each item was required to demonstrate a siyiificant item-total correlation (g <. 10) in order to remain part of the scale. Pearson product-moment correlations were exanined between each item raw score and the total raw score minus the item rc score of the item being examined, to determine the extent to which each item was measuring the same dimension as the whole scale. Only items sigiificætly correlated (fi <. 10) with the total raw score remained on the Preliminary Scale. The items remaining after these three reductions comprised the final versioi of the Human Figure Drawing: Features of Disturbance Scale (HFD: FDS). Rater reliability of the HFD: FDS. Rater reliability of the Final HFD: FDS was based on the total raw score (sum of the item raw scores for all three drawings). Intra-rat»' reli^ility was determined by examining the Pes'son product-moment correlation coefficient between the first time scoring and the second time scoring of 0 » 25 drawings scored twice by a single exaniner. Inter-rater reliability was determined by calculating a Pearson prodiKt-moment correlation coefficient between the total raw score obtained by rater one m é the total raw score obtained by rater two on the 25 drawings scored by two raters. Relationship of HFD: FDS to standardized measures of intelligence. Discriminant validity of the HFD: FDS was investigated by examining the relationship between the HFD: FDS and two stmdardized measures of intelligence for the normal subjects. Pearson product-moment correlation coefficients were calculated between the HFD: FDS total raw score and the MAT-SF total standard score, and between the HFD: FDS total raw score and the DAP: OSS total standard score. Each correlation coefficient was evaluated for statistical significance (g = .05) by comparing it to the appropriate table of siyiificant correlation values when testing a two-tailed hypothesis. 5 8 Ab\iitv of HFD: FDS to differentiate beW^een oroups. The HFDs of the exceptional group were rated using the Final Scoring Manual (Appendix C). Each of the 92 exceptional subjects was paired with a rwrmal subject based on matching age, race, and MAT-SF total standard score ( plus or minus one SEM which is e(^jal to six standard score points). The mean age ( in years) of the exceptional youp was 10.47 ( ^ = 1.43), while that of the matched normal youp was 10.43(50= 1.49); t (9 1 )= 1.07,a= .29. Each pair was matched precisely for race. The mean MAT-SF total standard score of the exceptional group was 88.70 ( ^ = 15.03), while that of the normal youp was 89.29 (SO = 13.94); t (9 1 )= 1.56,fi.= .12. Concurrent validity of the HFD: FDS was evaluated by comparing the mean HFD: FDS total raw scores rtîtained by these two groups to determine if these scores could significantly differentiate between the groups. The mean HFD: FDS total raw scores from the two groups were examined for significant difference (fi = .05) using a two-tailed t-test for paired observations. Additional HFD: FDS»>alvses. Several additional aspects of the scale were evaluated. First, the degree to which the mæ, woman, and self drawings reflect a common dimension (a type of alternate forms reliability) was investigated by examining the relationship among the individual draw ir^of the normal gwip. P e » ^ product-moment correlation coefficients were calculated between the total number of items present on each individual drawing and the total number of itans present on each of the two remaining individual cfe'awings. Pearson product-moment correlation coefficients were also calculated between the total number of items present on each individual A'awing and the combined total number of items present on the remaining two drawings. This same procedure was performed on the A'awings of the exceptional group. Additionally, a single factcr repealed- measures Analysis of Variance (ANOVA) was calculated to examine potential differences among the mean total number of itons present per individual drawing for the normal group. This procedure too, was performed on 5 9 the exceptional group. Second, potaitial race aid regioral differences were examined for the HFD: FDS. A two-factor factorial analysis of varioice ( ANOVA) was conducted to determine if the mean HFD: FDS total raw scores obtained by subjects from each of the four regions differed when compared to the mean HFD: FDS total raw scores obtained by subjects from each of the other regions; if the mean HFD: FDS total raw score obtained by black subjects differed from that obtained by white subjects; and if the mean HFD: FDS totel raw scores differed because of interaction between the factors of race and geographic region. Third, the clinical utility of the HFD: FDS, or its ability to predict group membership of individual subjects, was evaluated. A procedure proposed by Meehl and Rosen ( 1955) to determine diagnostic accuracy was followed to accomplish this. Diagnostic accuracy for identifying exceptionality was determined by dividing the number of subjects correctly classified as exceptional on the scale by the total number of subjects classified as exceptional on the scale. Diagnostic accuracy for identifying normality was determined by dividing the number of subjects correctly classified as normal on the scale by the total number of subjects classified as normal on the scale. Total accuracy equaled the total number of subjects correctly classified divided by the total n. A follow-up question is germane to this type of clasification analysis. Huberty ( 1984) recommended that a diagnostic accuracy rate, like that described above, should be evaluated to determine how much better one cai predict group membership based on the classification rule versus using some chance method of assignment. The "improvement-over-chaice index" (also termed the "reduction-in-error index") proposed by Huberty (p. 168) was used to make this determinatiwi. The formula for this index is as follows, where i is the reduction-in-error 5 0 I = ______l-He irelex, Ho is the observed hit rate, and Hg is the hit rate ©cpected by chance. Chapter IV Results As indicated in Chapter I, this study explored the development and usefulness of a Hum«i Figure Drawing (HFD) scoring system designed to idaitify potential emotional or behavioral disorders in children. This general issue subsumed a number of specific questions. Each specific research question is restated below, followed by the relevant results. Questions one, two. and three attt'essed the development of items for such a scale while questions four, five, and six evaluated the utility of tte resulting scale. Following the formal research questions, the results of a number of additional scale analyses are provided. Scale Development Prior to scoring the actual data collected for this study, each rater was required to pass a competency exam to assure the items w ^e being scored accurately. The competency exam was comprised of 10 drawings (four mwi, three woman, eraJ three self (Sawings) and satisfactory performawe was defined as yeater than 90 percent ay'eement between each rater's scoring and the correct scoring predetermirwd by the author and his advisoi'. Competency «cam results are presented in T^le Seven. Question one. Which of the HFD items reported in the literature as indicative of emotional or behavioral disturbance are ra'e or unusual (occjrring with less than 16% frequency) in the drawings of normal subjects? 61 6 2 T a b le ? Percenteoe of Item-Aoreement bv Raters on Compelencv Exam Number of Items Percentage of Rater Correct (of 940) Item Agreement A 900 96 B 911 97 C 905 96 D 882 94 After successfully completing the competency exam the four raters scored the drawings of the nwmal group to determine which items occurred infrequently anong normals. Item fretyjency distributions were constructed separately for the man, woman, and self drawings. Cut-off scores were established fir sach structural item by identifying the point on each frequency distribution where less than 16 percent of the normal group fell. These cut-off scores are presented in Table Eight The structural items were, thereafter, scored as one ( present) cr zero (absent) for each drawing. The drawings were also scored for the presence (one) or absence (zero) of each content item. A frequency distribution was constructed for each content item for the male, female, md self drawings. Tte mean occurrence rate of each content item, across all three (Sawings, was calculated and only content items with a mem occurrence rate of less thm 16 percent remained m the Preliminary Scale. Tte frequency of occurrence of each item, structural as well as content, is presented in Teri)le Nine. Based upon examination of Table Nine, 15 items were eliminated from the scale because they occurred too frequently anor^ normals. Sets of 6 3 T ab led Cut-off Scores for Structural Items on the Prelimlnav Scale Fiÿjre Dimensions and Ratios Item Man Woman Self 1 ) Tall Figure > 162 > 166 > 154 2) Short Figure <64 <65 <60 3) Large Figure > 14,450 > 13,950 > 12,000 4) Small Figure < 2,200 < 2,450 < 1,800 5) Top Placement < 0.30 <0.34 <0.33 6) Bottom Placement >0.62 >0.65 >0.67 7) Left Placement <0.33 <0.35 <0.38 8) Right Placement >0.51 >0.51 >0.51 10) Arm Length Urœven <0.75 <0.73 <0.73 11) Leg Length Uneven <0.86 <0.88 <0.87 12) Hand Area Uneven <0.57 < 0.55 <0.51 13) Foot Length Uneven <0.71 <0.75 <0.75 14) Torso/Staice Dispro < 0.25 or > 1.00 <0.25 or >0.96 < 0.27 cr > 0.90 17) Figure W/Ht Dispro <0.35 or >0.81 < 0.33 or > 0.88 <0.33 or >0.88 18)HeadW/Ht Dispro < 0.63 or > 1.23 < 0.60 or > 1.38 < 0.62 or > 1.33 19) H%d A/Fig A Dispro < 0.025 or > 0.27 <0.025 or >0.26 <0.025 or >0.24 20) Arm L/Leg L Dispro < 0.47 or > 1.34 <0.47 or >2.02 < 0.50 or > 1.35 21) Arm L/Fig Ht Dispro < 0.15 or > 0.38 < 0.13 or >0.37 <0.16or>0.41 22) Leg L/Fig Ht Dispro < 0.19 or > 0.47 <0.12or>0.43 < 0.22 or > 0.48 23) Hold A/Fig A Dispro <0.003 or >0.0225 <0.0025 or >0.0210 < 0.003 or> 0.0255 24) Foot L/Fig Ht Dispro <0.060 or >0.200 < 0.060 or >0.165 < 0.065 or > 0.200 Note: Items 1-4 are actual measurements ( mm). Items 5-24 are ratios of one measurement to another. See the Preliminary Scoring Manual (Appendix B) for detailed item descriptions. Items 9 ,15, and 16 are content items and are omitted from this table. 6 4 T a b le s on the Man. Woman, and Self Drawinos. and the Mean Occurrence Across All Three Orawinos Frequency Of Occurrence ( percent) Item Man Woman Self Mean DTall Fi^re 15 14 15 15* 2) Short Figure 15 15 15 15* 3) Large Figure 15 15 15 15* 4) Small Flÿjre 15 15 14 15* 5) Top Placement 15 15 15 15* 6) Bottom Placanent 15 15 15 15* 7) Left Placement 15 13 15 14* 6) Riÿit Placement 15 15 15 15* 9) Runs-Off Page 0 2 1 1* 10) Arm Length Uneven 12 15 15 14* 11) Leg Length Uneven 15 15 13 14* 12) Hand Area Uneven 15 15 15 15* 13) Foot Length Uneven 14 13 13 13» 14) Torso/Stance Disproportion 15 15 15 15* 15) Legs Drawn Together 9 7 11 9* 16) Baseline Drawn 9 7 8 8* 17) Figure W/Ht Disprcpjrtion 14 15 15 15* 18) Head W/Ht Disproportion 15 15 15 15* 19) Head A/Figure A Disproportion 14 15 11 13* 20) Arm L/Leg L Disproportion 15 15 15 15* 21 ) Arm L/Figure Ht Disproportion 15 15 15 15* 22) Leg L/Figure Ht Disproportion 15 12 15 14* 23) Hand A/Figure A Disproportion 15 13 15 14* 24) Foot L/Figure Ht Disproportion 13 15 15 14* 6 5 Table 9 (continued) Frec^jency Of Occurrence ( percent) Item Mm Woman Self Mean 25) Lettering/Numbering On Clothing 1 5 19 8 * 26) Lettering/Numbering Elsewhere 7 9 7 8 * 27) Page Upriÿit 100 100 99 100 28) Page Rotated 0 0 1 0* 29) Slanted Figure ( 15 D e g ^ ) 4 4 3 4* 30) Slfflited Flÿjre (25 Degrees ) 0 2 1 1* 31) Left-Facing Profile 4 4 2 3* 32) Right-Facing Profile 3 3 5 4* 33) Figure Facing Away 0 0 0 0* 34) Failed Inteyation Of Bod/ Po'ts 12 16 12 13* 35) Træspæmcies 9 8 4 7* 36) Restart 22 13 12 16 37) Head Omitted 0 0 0 0* 38) Hair Omitted 7 2 7 5» 39) Eyes (knitted 1 0 2 1* 40) Nose Omitted 6 6 11 8* 41) Mouth Omitted 2 2 4 3» 42) fteck Omitted 22 22 26 23 43) Torso Omitted 1 2 2 2* 44) Arms Omitted 2 2 2 2* 45) Fingers Omitted 11 12 12 12» 46) Legs Omitted 1 2 2 2* 47) Feet Omitted 9 9 4 7* 48) Clothing Omitted 12 13 16 14* 49) Face Erasure 32 25 31 29 50) Chœt Erasure 22 20 15 19 6 6 Table 9 (continued) Fretpjency Of Occurrence ( percent) Item M«i Woman Self Mean 51) Arm Erasure 25 21 23 23 52) Hand Erasure 20 12 12 15* 53) Crotch Erasure 8 9 7 8* 54) Leg Erasure 20 25 20 22 55) Foot Erasure 26 25 21 24 56) Outside Figure Erasure 23 18 23 21 57) Face Shading 4 3 5 4» 58) Chest Shading 21 22 28 24 59) Arm Shading 14 16 22 17 60) Hand Shading 1 2 3 2* 61) Crotch Shading 9 18 12 13* 62) Leg Shading 12 18 23 18 63) Foot Shading 7 9 10 9* 64) Outside Of Figure Shading 4 1 5 3* 65) Vacant Eyes 15 13 17 15* 66) Closed Eyes 0 0 0 0* 67) Crossed Eyes 2 2 1 2* 68) Eyes %z1ng Upward 1 1 1 1* 69) Eyes Gazing Downward 3 2 5 3* 70) Eyes Gazing Left 2 3 1 2* 71) Eyes Gazing Right 2 2 2 2* 72) Smiling Mouth 87 87 85 86 73) Frowning Mouth 1 1 0 1* 74) Straight Line/Slash Mouth 3 2 2 2* 75) Mouth With Teeth Visible 15 16 16 16* 76) Object In Mouth 5 1 1 2* 6 7 Table 9 (cOTtlnued) Frequency Of Occurrence (percent) Item Mm Woman Self Mem 77) Reaching Arms 1 2 2 2* 78) Outstretched Arms 28 31 27 29 79) Hanging Arms 53 49 54 52 80) Arms Pressed To Torso 12 13 II 12* 81 ) Arms Inconsistently Positioned 4 4 4 4* 82) Hands Cut-Off 8 6 7 7* 83) Hands Hidden 3 4 5 4* 84) Fists 2 1 2 2* 85) Talon-Like Fingers 5 6 8 6* 86) Aggressive Symbols 0 0 1 0* 87) Object Attached To Figure 7 7 5 6* 88) Background Filled-ln 8 8 13 10* 89) Broken Lines 16 12 10 13* 90) Buttons 13 10 7 10* 91) Non-Human/Monster 1 1 1 1* 92) Multiple Figures 0 0 0 0* 93) Nude Figure 3 3 3 3* 94) Soldier/Cowperson 1 0 1 1* * Denotes items occurring with less the 16% frequency other items were combined or collapsed into sirgle items because they were logically related and, even when combined, still occurred infrecpjently among normals. Specifically, items 29 and 30 wa-e combined to become Slojted Figure ( 15 detrees or greater), items 32 and 33 6 8 were combined to become Profile ( left « 1 right-facing), and items 68-71 were combined to become Eyes Gazing (any direction). This left 74 items remaining on the scale. Question two. Which of the remaining items can demonstrate an acceptable level of rater agreement (80%) under intra-rater and inter-rater conditions? The percentage of item-agreement under intra-rater æd inter-rater conditions is presented in Table Ten. Every item showed a satisfactory level of rater item-ayaement under both conditions (uniformly exceeding the predetermined 80 percent criterion), thus no items were eliminated because of poor rater item-agreement The number of items on the scale remained at 74. Table 10 Percentaoe of Intra-rater and inter-rater Item-Aireement for 25 Drawinos Percentage of Ayeement Item intra-rater Inter-rater 1 ) Tall Figure 100* 100* 2) Short Figure 100* 100* 3) Figure 100* 100* 4) Small Figure 100* 100* 5) Top Placement 96* 100* 6) Bottom Placement 100* 100* 7) Left Placement 100* 96* 8) Riÿit Placement 96* 92* 9) Runs-Off Page 100* 96* 10) Arm Length Uneven 96* 100* 11) Leg Lei^h Uneven 92* 96* 6 9 Table 10 (continued) Percentage of Agreement Itan Intra-rater Inter-rater 12) Hand Area Uneven 88» 96» 13) Foot Length Uneven 92» 96* 14) Torso/Stance Disproportion 96» 96» 15) Legs Drawn Together 100» 96* 16) Baseline Drawn 100» 100» 17) Figure W/Ht Disproportion 100» 100* 18) Head W/Ht Disproportion 96» 92* 19) Head A/Figure A Di^roporlion 92» 92» 20) Arm L/Leg L Disproportion 96» 92» 21 ) Arm L/Figure Ht Disproportion 96* 100* 22) Leg L/Figure Ht Di^roportion 100» 96» 23) Hand A/Figure A Disproportion 96» 96» 24) Foot L/Figure Ht Disproportion 96» 96» 25) Lettering/Numbering On Clothing 96* 100* 26) Lettering/Numbering Elsewhere 100* 100» 27) Page Rotated 100* 100* 28) Slanted F i^ re ( 15 Degrees) 96» 96» 29) Profile 100* 100* 30) Figure Facir^Away 100» 100» 31 ) Failed Intégration of Bod/ Parts 92» 92» 32) Transparencies 88» 96» 33) Head Omitted 100» 100* 34) Hair Omitted 100» 100» 35) Eyes Omitted 100» 100» 36) Nose Omitted 100» 100» 37) Mouth Omitted 100* 100* 7 0 Table 10 (continued) Percentage of Agreement Item Intra-rater Inter-rater 38) Torso Omitted 100» 100» 39) Arms Omitted 100* 100» 40) Fingers Omitted 100» 100* 41) Legs Omitted 100» 100* 42) Feet Omitted 100» 96» 43) Clothing Omitted 96* 88» 44) Hand Erasure 100* 96* 45) Crotch Erasure 100* 100* 46) Face Shading 92» 100» 47) FteKi Stradir^ 96» 100* 48) Crotch Shading 100» 80* 49) Foot Shading 92* 96* 50) Outside Of Figure Shadirg 100* 96* 51) Vacant Eyes 100* 100* 52) Closed Eyes 100* 100* 53) Crossed Eyes 100* 100* 54) Eyes Gazing 100* 100» 55) Frowning Mouth 100* 100* 56) Straight Line/Slash Mouth 100* 100* 57) Mouth With Teeth Visible 100* 100* 58) Object In Mouth 100» 100» 59) Reaching Arms 100* 96» 60) Arms Pressed To Torso 100* 100* 61) Arms Inconsistently Positioned 96* 96* 62) Hwds Cut-Off 92» 100» 63) Hands Hidden 100* 100* 7 ( Table 10 (continued) Percentage of Agreement Item intra-rater Inter-rater 64) Fists 100* 100* 65) Talon-Like Fingers 100* 92* 66) Aggressive Symbols 100* 100* 67) Object Attached To Figure 100* 100* 68) Background Filled-ln 92* 100* 60) Broken Lines 96* 92* 70) Buttons 100* 100* 71) Non-Human/Monster 100* 92* 72) Multiple Fiÿjres On Page 100* 96* 73) Nude Figure 100* 100* 74) Soldier/Cowperson 100* 100* Denotes satisfactory level of rater item-ayeement ( > 8085) Question three. To what degree is each remaining item related to the total group of remaining items? Item-total correlations were determined by computing a Pearson product-moment correlation coefficient between each item score aid the total scale score minus the item score being examined. These results are presented in Table Eleven. Forty-three items were deleted from the scale because of non-significant item-total correlations, leaving the total number of items on the scale at 31. These remaining items comprised the final version of the Human Figure Drawings; Features of Disturbwice Scale ( HFD: FDS), presented in Appendix C. 7 2 Table 11 Preliminary Scale Item-Total Correlations Item item I) Tall Figure .04 38) Torso Omitted .05 2) Short Figure .17* 39) Arms Omitted .19* 3) Large Figure .03 40) Firgers Omitted .23* 4) Small Figure .14* 41) Legs (knitted .09 5) Top Placement .06 42) Feet Omitted .24* 6) Bottom Placement -.07 43) Clothing Omitted .12* 7) Left Placement .25* 44) Hand Erasure -.19 8) Right Placement -.02 45) Crotch Erasure .01 9) Runs-Off Page .12* 46) Face Shading .07 to) Arm Length Uneven .08 47) Hold Shading .07 11) Leg Length Urœven .13* 48) Crotch Shading -.03 12)FtendArea Uneven -.08 49) Feet Shading .17* 13) Foot Length Uneven .28* 50) Outside Of Figure Shading .22* 14) Torso/Stance Disproportion .20* 51) Vacant Eyes .05 15) Legs Drawn Together .03 52) Closed Eyes .00 16) Baseline Drawn .22* 53) Crossed Eyes -.04 17) Fig W/Ht Disproportion .24* 54) Eyes Gazing -.01 18) Head W/Ht Disproportion .20* 55) Frowning Mouth .13* 19) itead A/F ig A Dispnprtion .01 56) Stralÿit Line/ 20) Arm L/Leg L Disproportion .18* Slash Mouth -.05 21) Arm L/Fig Ht Disproportion .25* 57) Mouth With Teeth Visible -.05 22) Leg L/Fig Ht Disproportion .19* 58) Object In Mouth .10 23) Hmd A/F ig A Disproportion .17* 59) Reaching Arms .16* 24) Foot L/Fig H Disproportion .05 60) Arms Preston To Torso .00 25) Lettering/Numbering 61 ) Arms Inconsistently On Clothing -.06 Positioned -.04 7 3 Table 11 (continued) Item r Item r 26) Lettering/Numbering 62) Hands Cut-Off .21* Elsewhere .08 63) Hads Hidden -.05 27) Page Rotated .17* 64) Fists .06 28) Slanted Flÿjre .06 65) Talon-Like Fingers .04 29) Profile .09 66) A g g ^iv e Symbols .14* 30) Figure Facing Away .00 67) Object Attached To Fiÿjre .20* 31 ) Failed Integration Of 68) Background Filled-ln .15* Body Pa'ts .27* 69) Broken Lines -.21 32) Transparencies .05 70) Buttons -.12 33) Head Omitted .00 71 ) Non-Human/Monster .03 34) Hair Omitted .20* 72) Multiple Figures .00 35) Eyes Omitted .11 73) Nude Figure .07 36) Nose Omitted .21* 74) Soldier/Oowperswi .06 37) Mouth Omitted .26* Denotes significant Item-total correlation (fi <. 10) Scale Utility Question four. Does the HFD; FDS have satisfactory Intra-rater md Inter-rater reliability? Rater-reliability was based on the total raw score of the final version of t1« HFD: FDS. A P e a ^ product-moment correlation coefficient of .94 (a < .0001 ) was obtained between the first time sxring aid second time scoring of 25 drawings scored twice by a single rater. A Pearson product-moment correlation coefficient of .9 1 (ê < .0001 ) was obtained between the 7 4 scoring of rater one and the scaring of rater two on 25 drawings scored by two different raters. Thus, the HFD: FDS demonstrated a high level of rater-reliability under intre-rater and inter-rater conditions. Question five. Cai the HFD: FDS be shown to diverge from intellectual ability measures by demonstrating low correlations with staidardized measures of intelligence? Pearson product-moment correlation coefficients were calculated between the HFD: FDS total raw score and the MAT - SF total standard score and between the HFD: FDS total raw score the DAP: QSS total standard score for the matctwd normal group (n = 92). Anrof -.19 (fi > .05) was obtained between the HFD: FDS and the MAT - SF, while an r of -.55 (fi < .05) was obtained between the HFD: FDS and the DAP: QSS. These results reveal the HFD: FDS was negatively correlated with intellectual or developmental factors as measured by both the MAT - SF and the DAP: QSS. The association between the HFD: FDS and MAT - SF was low and not statistically significant at fi > .05, indicating the HFD: FDS diverged from intelligence. The substantial and statistically simificant (fi< .05) association between the HFD: FDS and the DAP: QSS suggests these instruments are inversely related. The degree of this relationship was likely a function of items which were shared by the two scoring systems. For exar.ple, rou^ly one-third of the HFD: FDS items had scoring criteria which strongly resembled particular DAP: QSS items. Question six. Can the HFD: FDS differentiate between a group of children who placed in special settings for treatment of emotional or behavioral disorders and a ytwp of children who are enrolled in regular classrooms? The HFD: FDS total raw scores of the matched normal and exceptional groups were compared using a two-tailed i-test for paired observations. The results of this comparison appear in Table Twelve. The obtained t-value was significant (fi < .05), suggesting the scale did differentiate between youps of normal and emotionally or behaviorally disordered subjects. 7 5 Table 12 Means. Standard Deviations, and t-Test for HFD: FDS Group Differences @Dup n Mean SD t-test Normal 92 9.16 5.92 4.82* Exceptional 92 13.47 7.33 *fi<.0001 Additional HFD: FDS aialvses. Several additional aspects of the scale were evaluated. First, the degree to which the man, woman, and self drawings reflect a common dimension (a type of alternate forms reliability) was investigated by examining the relationship anong these individual (Sawings of the normal group. P e a ^ product-manent correlation coefficients were calculated between the total number of items present on each individual drawing and the total number of items present on each of the two remaining individual drawings. Pea'son product-manent correlation coefficients were also calculated between the total number of items present on each individual drawing and the combined total number of items present on tte remaining two drawings. Table 13 reveals m odale to substantial associations between each individual clawing and each of tte othæ two drawings, md between each drawing md the other two drawings combined. This same procedure was performed on tte exceptional group and tte results appear in Table 14. Again, moderate to substaitial associations between each individual drawing axJ each of tte other two ^'swings, aid between each drawing and tte other two drawings combined were found. 7 6 Table 13 Corrélation Matrix of HFD: FDS Scores Between Each Individual Drawing, and Between Each Individual Drawing and the Remaining Two Drawings ( Normals, n = 186) Drawing Man Woman Self Woman .52 --- Self .42 .46 --- M n + Wonan ------.50 Man + Self --- .58 --- Woman + Self .55 ------ ftote; All r ‘s signifient at b < .001 Table 14 Correlation Matrix of HFD: FDS Sores Between Each Individual Drawing, end Between Each Individual Drawing and the Remainlno Two Drawinos ( Exceotionals. n= 92) Drawing Man Woman Self Womn .54 ------Self .50 .54 --- Man + Woman ------.59 Man + Self --- .63 --- Woman + Self .59 ------ Note; All r ’s are significant at fi < .001 7 7 Additionally, a single factor repeated-measures Analysis of Variance (ANOVA) was calculated to examine potential differences among the mean total numt)er of items present per individual drawing for the normal group. These results are presented in Table 15. The obtained F-value was not significant (g > .05), indicating the mean total number of items present per individual (Rawing did not differ beyond chance fluctuation. The same procedure was performed on the exceptional group. These results a'e presented in Table 16. Again, the obtained F-value was not signifient (& > .05), indicating the mean total number of items present per individual drawing did not differ beyond chance fluctuation. Table 15 Means. Standard Deviations, and F-Test for Total Number of Items Præent per Drawing (Normals. n=186) Drawing n Men SO F-test Man 186 2.82 2.36 0.479 Woman 186 2.74 2.24 Self 186 2.92 2.56 Note; F-test was not significant (fi > .05) 7 8 Table 16 Means. Standard Deviations, and F-Test for Total Number of Items Present per Drawing (Exceotionals. n= 92) Drawing ji Mean SD F-test Man 92 4.54 3.12 0.07 Woman 92 4.49 2.97 Self 92 4.44 2.77 Note: F-test was not significant (fi > .05) Second, potential race and regional differences were examined for the HFD: FDS. A two-factor factorial analysis of variance (ANOVA) was conducted to determine if the mean HFD: FDS total raw scores obtained by subjects from each of the four regions differed when compared to the mean HFD: FDS total raw scores obtained by subjects from each of the other regions; if the mean HFD: FDS total raw score obtained by black subjects differed from that obtained by white subjects; md if the mean HFD: FDS total raw scores differed because of interaction between the factors of race and geographic region. These results are presented in Table 17. The obtained F-values were not significant (fi > .05) regardif^ the comparison of race, geographic region, and interaction between these two factors, indicatif^ these factors did not influence the mean HFD: FDS total raw score beyond chmce fluctuation. 7 9 Table 17 mSSmS. Standard Deviations, and F-Tests for HFD: FDS Race/Reolonal Differences Region Northeast North Central South West F-Tests Race White n 36 43 46 30 X region = 1.46 7.8 6.7 9.9 7.9 X race = 0.79 6.5 4.8 6.4 4.4 interaction = 0.74 Block n 2 2 20 7 M 12.5 4.5 10.3 10.7 5.0 0.7 5.4 6.2 Note; F-tests were not significant > .05) Third, the clinical utility of the HFD: FDS, or its ability to predict g ’oup membership of individual subjects, was evaluated. A procedure proposed by Meehl and Rosen ( 1955) to determine diagnostic accuracy wss followed to accomplish this. Diagnostic accuracy fcr identifying exceptionality was determined by dividing the number of subjects correctly classified as exceptional on the HFD: FDS by the total number of subjects classified as exceptional on the HFD: FDS. Diaywstic accuracy for identifying normality was déterminai by dividing the number of subjects correctly classified as normal on the HFD: FDS by the total number of subjects classified as normal on the HFD: FDS. Total accuracy equaled the total number of subjects ctrrectly classified divided by the total n. Several potential cut-off scores were considered aid 11 was found to be the optimal cut-off score. When using 11 as a cut-off score, the diagnostic accuracy for exceptionality was .67, the diagnostic accuracy for 8 0 normality was .65, and the total diagnostic accuracy for the HFD: FDS was .66. Thus, based on these results the HFD: FDS correctly classified behavior disordered subjects versus non-behavior disordered subjects with 66 percent accuracy, or two-thirds of the time. The degree to which this level of diagnostic accuracy was an improvement over chance assiÿiment was determined by using the "improvement-over-chance index" (also termed the "reduction-in-error index") proposed by Huterty ( 1984 p. 168). The formula for ü^is index is as follows, where i is the reduction-in-error index, Hq is the observed hit rate, and Ho-He 1-He % is the hit rats expected by chance. This classification aialysis indicated that 32 percent fewer errors were made using the HFD: FDS than would be expected by choice. Chapter V Discussion This chapter opsns with a brief summary to review the major findings of the present investigation. This is followed by the int^pretation and discussion of these findings in the context of previous Human Figure Drawing reseaxrfi. The chapter closes by outlining the stud/s limitations and considering its implication for future r e s e a t on the HFD technique. Summary of the Study This study explored the development and usefulness of a Humai Figure Drawing (HFD) scoring system designed to identify potential emotional or behavioral disorders in seven to twelve year old males. Constructiai of this scale, termed the Human Figure Drawing: Features of Disturbance Scale (HFD: FDS), was based on both rational and empirical grounds. A pool of 94 preliminary items was selected from the relevant literature and from previously available projective HFD tests which were simil»' in nature. Each item was carefully defined to m *e its scoring as objective as possible. A Prelimiiwy Scoring Manual (Appendix B) was constructed which described each item's scoring criteria. Raters, who w«'e trained to use the Preliminary Scoring Manual, scored the drawings from a group of 186 normal children. The technical properties (actuarial valœ among normal children, percentage of item-agreement under intra-rater and inter-rater conditions, and item-total correlation) of eadi item were examined »id the number of items was reAiced based upon these properties. The remaining list of 31 items comprised the HFD: FDS (Appendix C). 81 8 2 The scale’s utility was evaluated on three major fronts. First, the rater reliability of the HFD: FDS was evaluated under intra-rater aid inter-rater conditions. The scale demonstrated satisfactory results in this respect, with intra-rater and inter-rater reliability coefficients of .94 and .91, respectively. Second, the scale's discriminant validity was determined by examining its relationship to standardized measures of intellectual ability. The HFD: FDS was not significantly related to the Matrix Analogies Test - Short Form ( MAT - SF) (Naglieri, 1985); but a significant, inverse relationship was found between the HFD: FDS and the Draw A Person: A Quantitative Scoring System ( Naglieri, 1988). Third, the HFD: FDS's concurrent validity was evaluated regarding its ability to distinguish between carefully selected criterion groups. HFD: FDS scores were found to significantly differentiate between groups of normal and disturbed children who were matched for age, sex, race, and level of intellectual ability. The HFD: FDS's diagnostic hit rate was determined to be .66, indicating the scale correctly identified individual subjects as normal or disturbed 66 percent of the time. This was shown to be a 32 percent reduction in classification errors as compared chance assignment Discussion The first question asked by this study addressed the issue of how many itans which have been identified in the HFD literature as indicative of emotional or behavicral disturbance apf»ared rarely on the HFD's of normal subjects. Of the 94 preliminary items identified from the literature and previous tests, only 74 were found to occur with less thai 16 percent frequency among normal children. Thus, roughly 2 1 percent of the items which have been traditionally used as indicators of emotional or behavioral disturbance appe^ed with some regularity among normals. For example, of the 16 items related to erasure or shading of a particular body part ( believed to be reflective of a specific conflict or anxiety [ Machover, 8 3 1949] ), nine occurred commonly on the HFDs of normals ( mean rate of occurrence of 22 percent). This finding supported the suggestions of Koppitz ( 1968), Naumberg ( 1953 [cited in Falk, 1981 ] ), and Ogdon ( 1977) regarding the presence of such items on the HFD’s of normal as well as disturbed subjects. This finding may help explain the inconsistency and lack of support apparent in the relevant literature to date, as such items may have confounded results. It follows that the ^ ility of a global rating procedure whose aim is to differentiate between normal and disturbed subjects may be enhanced by the exclusion of these frequently occurring items. The second question addressed by this study examined the degree to which the projective items that appeared Infrequently among normals could be consistently scored. Each of the 74 items identified in Question one was found to possess a satisfactory level (greater than 80 percent) of rater itan-agreement under intra-rater and inter-rater conditions. This was a general improvement over rater item-agreement levels revealed in the early HFD literature, but it was fairly consistent with more recently reported findings. As noted in the literature review, recent improvements on this dimension likely reflected the use of more detailed scoring instructions and more specific training procedures being provided for raters. Both of these features were present in the current study and the present results, at least indirectly , supported earlier finding that explicit scoring instructions and more specific rater-training procedures enhance the consistency of HFD rating (Cassel, Johnson, & Burns, 1958; Strumpfer, 1963). The study's third (fjestion explored the feasibility that a meaningful portion of items identified in Questions one and two could te shown to reflect a common dimension, and that the number of these items would be sufficient to constitute a useful scale. Of tte 74 items identified in (testions one and two, 3 1 were found to significantly relate to the total item- youp ( i.e., significant item-total correlations were found). These results demonstrated there 8 4 Is some degree of homogeneity emong HFD: FOS Items and, thus, the scale as a whole tends to be unidimensional rather than multidimensional. The number of items reflecting this common dimension was sufficient to constitute a useful scale, as shown by the resulting scale's ^ ility to differentiate between criterion-groups (see discussion of Question six, below). Further reflective of the HFD: FDS's homogeneity were findings demonstrating that each individual drawing was significantly correlated with each of the other drawings, and with the other two drawings combined. A related issue should be pointed out at this stage. It must be noted that differential treatment of the man versus the woman versus the self drawing was hypothesized ( but not consistently supported) in past literature to reflect differences in a subject's attitude tow»'d males and females or the subject's feelings about self (e.g., Machover, 1949). The present results showed no differential treatment of individual drawings for the normal or the exceptiorel group. It must be noted, however, that the aAninistration procedure (Appendix A) in the present study prescr Ibad which sex was to be drawn first, thereby preventing consideration of sex of first-drawn figure as an item. While the present findings did not rule-out the potential for differential treatment of individual drawings being a useful item, it did not supptrt the hypothesis that it is. This is a tqiic which merits further exploration, particularly under different administration conditions. Question four focused on tte evaluation of rater reliability for the HFD: FDS. Reliability coefficients of .94 end .91 were obtained under intra-rater and inter-rater conditions, respectively. Such findings are higher than those revealed in the literature, even higher than the reliability coefficients cited for recent studies which have tended to use more explicit scoring criteria and improved rater-training procedures. Thus, ü » HFD: FDS's rater reliability was excellent These results, as did the earlier findings on inter-rater item- agreement, at least indirectly supported tte contention that greater detailed scoring instructions æd more ^lecific rater-training procedures enhance the consistency of HFD 8 5 rating (Cassel, Johnson, & Burns, l958;Strumpfar, 1963). The stud/'s fifth question was designed to evaluate the divergent or discriminant validity of the HFD: FDS. The relationship(s) between the HFD: FDS and two measures of intellectual ability were examined. A low, non-significant negative correlation was found between the HFD; FDS and the MAT - SF (Naglieri, 1985), suggesting the scores on these two tests reflected diverging operations. However, a substantial and statistically significant negative correlation was obtained between the HFD: FDS and the DAP: QSS ( Naglieri, 1988), suggesting these instruments were inversely related. The degree of this latter relationship was likely a function of items which were shared by the two scoring systems. For example, roughly one-third of the HFD: FDS items had scoring criteria which strongly resembled particular DAP: QSS items. The relationship was negative because the s h » ^ items were keyed differently between the two scoring systems. For example, omission of pa'ticular body po'ts was associated with a lower level intellectual functioning and a hidier probability of emotional or behavioral disturbance. These results indicated the HFD: FDS did diverge from intellectual ability as measured by the MAT - SF. This finding was somewhat of a contrast to the results of prior factor- analytic investigations (Adler, 1970; Nichols & Strumpfer, 1962) which suggested HFD performance essentially reflected cognitive maturity or sophistication. However, it may be that the HFD: FDS would be associated with one of the less powerful factors which were a p p o r t in the factor-malytic findings, as it was suggested these factors may have had some relationship to personality variables (Adler, 1970). The reduction of prelimirwy HFD: FDS items on the basis of non-significant item-total correlations may have eliminated items which were mere associated with cognitive matur ity, allowing the HFD: FDS to load on one of these lesser factors. This is a speculative assertion, but may serve as a focus for future resea'ch. 8 6 Question six evalusted the concurrent validity of the HFD: FDS by examining its ability to differentiate between a group of diiidren who were di^noeed as having emotional or behavioral disorders and a group of children who were enrolled in regular classrooms. The scale significantly differentiated between these groups. This differentiation between gross levels of psychological adjustment by a global HFD measure was consistent with the results of the majority of past investigations on that topic, as reviewed Ketfiill ( 1984), Roback ( 1968), and Swensen( 1957, 1968). A follow-up issue was evaluated regarding the HFD: FDS's diagnostic accuracy in predicting the status of individual subjects on the criterion variable. It was datermined that the scale correctly classified children as disturbed versus non-disturbed 66 percent cf tie time. Furtter analyses demonstrated that the HFD: FDS provided a 32 percent reduction in classification errors as compared to classification by chance. Data of this type was not directly found in the existing HFD literature. However, the raw classification data was provided by Koppitz ( 1968, p. 42) for a similar study investigating the usefulness of her Emotional Indicators in differentiating between children who were patients at a guidance clinic and children who were well-adjusted Koppitz similarly determined optimal cut-off score for her system and found that the system did differmtiate between groups. The current author applied the same procedures used in the present study ( Huberty, 1984) to the Koppitz data (using Koppitz’ chosen optimal cut-off score for her system) to determine the diagnostic accuracy of the Emotional Indicators. The Koppitz syslKn orrectly classified children as clinic patients versus well-adjusted 84 percent of the time. Further malyses demonstrated that her system provided a 68 percent reduction in classification errw s as compæed to classification by chance. Thus, while both systems provided a classification method appreciably better than chance, these results suggest that Koppitz’ system more accurately identified disturbance versus non-disturbance than the 8 7 HFD: FDS. However, differences between the scoring systems themselves, between the methodological ^preaches used, and among the subjects studied must be considered when comparing the remits of these two studies. While the total number of items appearing on the HFD: FDS and the Koppitz scale was similar wkJ mffliy of the items had similo' names, the actual scoring criteria for items differed markedly. This was likely due, at least in part, to differences in the manner in which the scales were developed. For example, Koppitz a rb itra ily defined cut-off scores for items related to size based on her experience with drawings, rather ths) by precise measurement as with the HFD: FDS. Also, no attempt to provide homogeneity among items was made in the development of Koppitz' system. Thus, the surface similarities between these two scales can be misleading as to the scales' actual content Methodological contrasts between the two studies may also lend to better understanding of the differing results. For example, the subjects in the Koppitz ( 1968) study were not matched for age, race, and intellectual level as in the HFD: FDS study. The Koppitz study included male and female subjects while the present investigation used only males. Also, subject scores were based on one HFD per subject in the Koppitz study as opposed to three HFDs used in the present investigation. It is also likely that variation among the groups studied further COTifounded the comparison of the two projects. For example, it is not known how similar the guidance clinic patients were to diil(Nn who were identified in the present study as having an emotional or behavioral disorder. Thus, while the HFD: FDS did not classify subjects as accurately as did the Koppitz system, the results of the two studies may not be directly comparable. Investigations to more directly compare the two systems would provide useful information. 8 8 Oyerall, the results of the present investigation were essentially consistent with projective HFD theory and research as it presently stands. Scorers consistently rated HFDs under conditions where explicit scoring criteria and adequate rater-training were provided. And, g1(Aa1 HFD characteristics were shown to be significantly related to gross levels of psychological adjustmwit On the practical side of matters the results indicated that the HFD: FDS may be useful as a screening measure for emotional or behavioral disorders in children as it correctly classified subjects appreciably better than chance. Despite the evidence supfxirting the HFD: FDS as a screening measure for emotional or betwvioral disorders in children, clinicians should not lose sight of the supplementary, less quantifiable uses of HFDs in general ( including the HFD: FDS). As Klopfer and Taulbee ( 1976) suggested, the technique miÿit be best regarded as a kind of yaphic behavior that takes cm particular meaning only as It is discussed with the subject and when it is viewed in the context of other information. Limitations of the Study A number of limitations are apparent regaxling the generalization of the present results. The study controlled for several subject characteristics (sex, age, race, and intellectual level) to enhaice the power of the group comparison techniques. Thus, the results of the present investigation cannot be viewed as relevait to females or to males who differ significantly from the characteristics of the subjects studied for this investigation. The particular setting(s) in which the data was collected placed further restrictions regarding genaralizafaility. While the carefully described diagnostic criteria used to select the exceptional group yielded a fairly homogeneous group for comparison to normals, it prevents the generalization of the presait results to groups which may be identified as exceptional by other criteria Othar limitations created by the study's desiyi include the elimination of sax- 8 9 of-first-drawn-figure as a possible scale item due to the administration instructions used, and failure to control for differing degrees of artistic ability among subjects. Imolications for Future Research Suggestions for future research appeared throuÿnut Vhè discussion above, but for improved clarity, they will be briefly reviewed here. First, other global HFD rating scales should be evaluated to determine if their quality can be enhanced by the elimination of items which occur frequently among normals. Second, the homogeneity of other global rating scales should be explwed by examining the relationship between each item and the total group of items, and between the individual man, woman, aid self drawings. In turn, should the scale’s level of homogeneity increase based upon the elimination of non-related items, the scale’s relationship with cognitive maturity should be reevaluated. Last, the direct comparison of the HFD: FDS to similar scales ( e.g., Koppitz’ [ 1968] system of emotional indicators) may help identify the most effective HFD screening method for particular populations. Appendix A HFD: FDS ADMINISTRATION INSTRUCTIONS * The following instructions should be read as printed and time limits are to be carefully enforced. Draw-A-Person Test: I'd like you to draw some pictures for me. You will have five minutes to do each (fr^awing. On the first blank page I 'd like the boys to draw a picture of a man md the girls to *'aw a picture of a women. Boys draw a picture of a mw and girls draw a picture of a woman. Make the very best picture you can. Të:e your time and work very carefully. Be sure to the whole person. Please begin. (Allow five minutes). This time, on the second blaik page, I wait Vie boys to draw a picture of a wanai and the girls to draw a picture of a man. Boys draw a picture of a woman aid girls draw a picture of a man. Mace the very best picture you cai. Tace your time and work very carefully. Be sure to draw the whole pa'son. Please begin. (Allow five minutes). Now, on the last blaik page. I'd like you to draw a picture of yourself. Again, draw the very best picture you can. Take your time and work very carefully. Please begin. (Allow five minutes). 9 0 Appendix 6 PRELIMINARY SCORING MANUAL Figure Height Figure heiÿit equals the distmce between the uppermost point of the figure and the lowermost point of the figure. Articles of clothing such as hats or shoes are Included in the measurement* while other objects (i.e., handbag, briefcase, backpack, baseball bat, etc.) are not included in the measurement Cut-off scores will be established to determine Tall Figure ( Item 1 ) and Shcrt Figure ( Item 2). Figure Area: Figure area equals figure heiÿit multiplied by figure width. Figure width equals the distance between extreme left and right points of figure. As with figure height, articles of clothing are Included in the measuranent of figure width while other objects are not included in the measurement Cut-off scores will be est^lished to determine Large Figure ( Item 3) and Small Figure ( Item 4). Vertical Placement: Vertical figure placement equals (CFH/2) + (ID)) / PH, where FH = figure height, ID = distance from top of figure to top edge of page, and PH = the height of the page (279 mm). Cut-off scores will be established to determine Top Placement (Item 5) and Bottom Placement ( Item 6). Horizontal Placement Horizontal figure placement equals ((FW/2) + (LD)) / PW, where FW = figure width, LD = distance from leftmost point of figure to left edge of p ^ , and PW 3 the width of the page (216 mm). Cut-off scores will be established to determine Left Placement ( Item 7) and Right Placement ( Item 8). Runs-Off Page ( Item 9): Score 1 if any portion of the figure touches any edge of the paper or runs-off the page, otherwise so re 0. * All measurements are to be made in millimeters. 91 9 2 Arm Length: Measure right and left arm lengths. Arm length equals distance between point where arm attaches to the body ( from arm pit) and furthermost point of arm, hand, or finger. If arm is bent at the elbow, arm length equals distance between point where arm attaches to body and elbow plus distance between elbow and furthermost point of arm, hand, or finger. See Figure One below for measurement of arm length. The length of the shorter arm will be divided by the length of the longer arm and cut-off scores will be established to determine Arm Length Uneven (Item 10). Arm Length Fiÿjre 1 ^rm Length Measurement Leg Length: Measure right and left leg lengths. Leg length equals distance between point where leg attaches to body ( i.e., crotch) and the furthermost point of leg, heel of foot or shoe, or toe. If leg is bent at the knee, leg length equals distance between point where leg attaüies to tody and knee plus distance between knee md furtha'mcst point of leg, heel of foot or shoe, or toe. See Figure Two below for measuranent of leg length. The length of the shorter leg will be divided by the length of the longer leg and cut-off scores will be established to determine Leg Length Unevm (Item II). 9 3 item 11 (continued) T T Leg Length Leg Length Leg Length Figure 2 Leg Length Measurement itendArea: Hand aree equals hand height multiplied by hand width. Hand height equals distmce t)etween uppermost point of hand ( including fingers) aid lowermost point of haid ( including fingers) without regard to orientation of hand. Hand width equal distance tietween extreme left and right points of hand ( including fingers) without regard to orientation of hand. See Figures Three and Four below for measurement of hand heiÿit and hand width. The smaller haral area will be divided by the larger hand area aid cut-off scares will tie established to determine Hoid Area Uneven ( Item 12). Hand Height Figure 3 Hand Height Measurement 9 4 Item 12 (continued) OiO Hand Vidth Figure 4 Hand Width Meesurement Foot Length: Foot length equals distance between rear of heel and end of toe or toe-end of shoe. If heel is not shown loot Iwigth equals distance between ankle and end of toe or toe-md of shoe. See Figure Five below for measurement of foot length. The shorter foot length will be divided by the longer foot length and cut-off so res will be established to determlre Foot Length Uneven ( Item 13). IL Foot Length Figures Torso/Stance Widths: The torso measurement is taken at the narrowest portion of the the torso (below neck, If present, and above crotch or point who^ legs attach). Stance width equals the distance between left-most point of foot (or bottom-most portion of leg If foot not present) and r # t most point of foot (or bottom-most portion of leg If foot not present). See F Igures Six and Seven below for measurement of the width of the narrowest portiwi of the torso and the stance width. Torso width will be divided by stance width and cut-off scores will be established to determine Torso/Stance Disproportion (Item 14). 9 5 Item H (continued) Vidth of Narrowest Portion of Torso Figures Width of Nyrowest Portion of Torso Measurement I I ) len 0 stance width Fiÿjre? Stance Width Meesurement Leos Drawn Together ( Item 15): Score 1 if legs are drawn together with no visible space between legs or if only one is leg visible in profile. Otherwise score 0. Baseline Drawn ( Item 16): Score 1 if a baseline is drawn (i.e. ground, yass) for figure to stand on. Otherwise score 0. Figure Width/Heicht: Figure width divided by figure height Cut-off scores will be established to determine Figure Width/Height Disproportion ( Item 17). 9 6 HeadWldth/Heloht Head width dlvklBd by head height Head width equals the distance between extreme left and right points of head, following the contour of the face. Measurement does not include ears, hair, or hat Head height equals the distance between the lowermost point of head and the uppermost point of head ( including hair or kiard but not neck). If figure is wearing a hat or a piled-high hair style, measure to the top of hat or hair unless the actual top of head is cleerlv visible. If top of heed is cksrly visible, measure only to that point See Figures Eight and Nine below for measurement of the width and height of hMd. Cut-off scores will be established to determine Head Width/Height Disproportion (item 18). Head Vidth Figures Heed Width Measurement Head Height T X A Figure 9 Head Height Meesurement 97 Head Area/Flaure Area Head area divided by figure area Heed area equals head width multiplied by head height Cut-off scores will be established to determine Head Area/Figure Area Disproportion ( Item 19). Arm length/Leo Lenoth: Average arm length divided by average leg length. Average arm length equals tlwsumof both arm lengths divided by two. Average leg length equals the sum of both leg lengths divided by two. Cut-off scores will be established to determine Arm Length/Leg Length Disproportion (Item 20). Arm Length/FioureHeitfit: Average o^m length divided by figure height Cut-off scores will be established to determine Arm Length/Figure Height Disproportion ( Item 21 ). Leg Length/Figure Height Average leg length divided by figure height Cut-off scores will l« est^lished to determine Leg Length/Figure Height Disproportion ( Item 22). Hand Area/Figure Area Average hand area divided by figure area Average hand area equals the sum of both hand areas divided by two. Cut-off scores will be established to detK-mine Hand Area/Figure Area Disproportion (Item 23). Foot Length/Figure Height Average foot length divided by figure height Average foot length equals the sum of both foot lengths divided by two. Cut-off scores will be established to determine Foot Length/Figure Height Disproportion (Item 24). Letterino/Numberino On Clothing or Figure ( Item 25): Score 1 if letters, words, phrases, or numbers appear on clothing or figure. Otherwise score 0. Letterino/Numberino Elsewhere ( Item 26): Score 1 if letters, words, phrases, or numbers appear elsewhere on the page. Otherwise score 0. PaoB Upright ( Item 27): Score 1 if figure is drawn with the diortest dimension of page (216 mm) at the top/bottom. Otherwise score 0. 9 8 Paœ Rotated ( Item 28): Score 1 1f figure Is drawn with the longest dimension of page ( 279 mm) at the top/bottom. Otherwise score 0. Slanting Figure ( 15 Degrees) ( Item 29): Score 1 if vertical axis of figure ( i.e., line from mid0oint of head witfth to midpoint of stance width) deviates by 15 degrees or more ( but less than 25 degrees) from a line perpendiculw* to bottom edge of page. Otherwise score 0. This item is measured by placing a clear template (see Appendix 0) over figure which shows horizontal and vertical axes, and t S and 25 degree deviations. Slanting Fioure ( 25 Decrees) ( Item 30): Score 1 if vertical axis of figure ( i.e., line from midpoint of head width to midpoint of stance width) deviates by 25 degrees or more from a line perpendicular to bottom edge of page. Otherwise score 0. This item is measured by placing a clear template (see Appendix D) over figure which shows horizontal and vertical axes, and 15 and 25 degree deviations. Left-Facino Profile ( Item 31 >. Score 1 if entire figure or head only is in left-facing profile. Otha^ise score 0. Ric^t-Facing Profile ( Item 32): Score I if entire figure or head only is in right-facing profile. Otherwise score 0. Figure Facing Awav From Viewer ( Item 33): Score 1 if entire figure or head only is facing away from viewer. Otherwise score 0. Failed Intecration of Body Parts ( Item 34): Score 1 for failure of anv portion of the following: Head attached to neck or top of torso, two arms (one if in profile) attached to the top half of the torso (above the half-way mark in the vertical measurement of the torso), and two legs (one if in profile) attached at the bottom of the torso. Otherwise score 0. Transparencies ( Item 35): Score 1 if any body pa*t shows through clothing (or another bod/ part). Otherwise score 0. 9 9 Restart (item 56): Score 1 1f one or more figures are obviously abandoned (m a/be erased, scratched-out, or merely left Incomplete) In addition to a more complete figure on the page. Otherwise score 0. Head Omitted ( Item 37): Score 1 if the figure’s head Is absent, otherwise score 0. Any attempted representation of a head cannot be scored as an omission. tteirDmitteddtem 38): Score 1 (per man, woman, and self drawing) if the figure has no hair (on head), o th ^ is e score 0. Any attempted representation of hair (on head) cannot be scored as an omission. Eves Omitted ( Item 39): Score 1 if the figure's eyes are absent, otherwise score 0. Any attempted representation of eyes ( including only one eye) cannot be scored as an omission. Nose Omitted (Item 40): Score 1 if the figure's nose is absent, otherwise score 0. Any attempted representation of a nose camot be scored as an omission. Mouth Omitted ( Item 4 1 ): Score 1 if the figure's mouth is absent, otherwise score 0. Any attempted representation of a mouth cannot be scored as an omission. Neck Omitted (Item 42): Score 1 if the figure's neck is absent, otherwise s o re 0. Any attempted representation of a neck cannot be scored as an omission. Torso Omitted ( Item 43): Score 1 if the fiÿjre's torso is absent, otherwise score 0. Any attempted representation of a torso cannot be scored as an omission. Arms Omitted ( Item 44): Score 1 if the figure has no arms, oUierwlse score 0. Any attempted representatim of wms ( including only one arm) cannot be scored as an omissim. Fingers Omitted ( Item 45): Score 1 if the figure has no fingers, otherwise score 0. Any attempted representation of fingers cannot be scored as an omission. 100 leos Omitted (Item 46): Score 1 1f the figure has no legs, otherwise score 0. Any attempted representation of legs ( including only one leg) carwuiit be scored as an omission. Feet Omitted ( Item 47): Score 1 if the figure has no feet, otherwise score 0. Any attempted representatlœi of feet ( including only one foot) cannot be scored as an omission. Clothing Omitted ( Item 48): Score 1 if the figure has no indication of clothing, otherwise score 0. Any attempted representation of clothes cannot be scored as an omission. Face Erasure ( Item 49): Score 1 if erasure is apparent in the ares of the figure's face, otherwise score 0. Chest Erasure ( Item 50): Score 1 if erasure is apparent In the area of the figure’s chest, otherwise score 0. Arms Erasure ( Item 51 ): Score I if erasure is apparent in the area of the figure's arm(s), otterwise score 0. Hand Erasure ( Item 52); Score I if erasure is apparent in the area of the figure's handCs), otherwise score 0. Crotch Erasure ( Item 53): Score 1 if erasure is apparent in the area of the figure's crotch, othwwise score 0. Legs Erasure ( Item 54): Score 1 if erasure is apparent in the of the figure's leg(s), otherwise score 0. Feet Erasure ( Item 55): Score 1 if erasure is apptrent in the area of the figure's feel (foot), otherwise score 0. Outside Of Figure Erasure (Item 56): Score 1 if erasure is apparent in ai area outside of the figure, otherwise score 0. toi FacsShsdi^: ( Item 57): Sore 1 1f pencil strokes ere present on tbe figu'^’s face which are designed to fill in an area as in coloring or darkening. Otherwise score 0. Chest Shading ( Item 58): Score I if pencil strokes are present on the figure's chest which are designed to fill in an area as in coloring or darkening ( including stripes or checks on clothing). Otherwise score 0. Lettering or numbering is not scored as shading Arm Shading ( Item 59): Score I if pencil strokes are present on the figure's arm(s) which are designed to fill in m area as in coloring or darkening ( Including stripes or checks on clothing). Otherwise score 0. Lettering or numbering jsngt scored as shadina Hand Shading ( Item 60): Score I if pencil strokes are present on the figure's hand( s) which are designed tc fill in a i area as in coloring or darkening. Otherwise score 0. Crotch Shading ( Item 61 ): Score I if pencil strokes are present on the figure's crotch area which are designed to fill in an area as in coloring or darkening ( including stripes or checks on clothing). Otherwise score 0. Leg Shading ( Item 62): Score I if pencil strokes are present on the figure's leg(s) which are desiyied to fill in wi area as in coloring or darkening ( including stripes or checks on clothing). Otherwise score 0. Letteriro or numbering Is not sored es shading. Feet Shading ( Item 63): Score I if pencil strokes are present on the figure's feet (foot) which are deigned to fill in an area as in coloring or darkening. Otherwise score 0. Outside Of Fioure Shading ( Item 64): Score 1 if pencil strokes are present outside of the figure which ^ desigied to fill in an area as in coloring or darkening. Otherwise score 0. Vacant Eves ( Item 65): Score 1 if figure's eyes are vacant ( i.e., open circles), otherwise score 0. Closed Eves ( Item 66): Score I if figure's eyes are closed, otherwise score 0. 102 Crossed Eves (Item 67): Score 1 if figure's eyes are crossed, otherwise score 0. Eves Oezino Upward ( Item 68): Score 1 if figure's eyes are gazing upward, otherwise score 0. Eves Gazing Downward ( Item 69): Score 1 if figure's eyes are gezing downward, otherwise score 0. Eves Gazino Left ( Item 70): Score 1 if figure's eyes are gazing toward the examiner's left, otherwise score 0. EvesOazino Right ( Item 71 ): Score 1 if figure's eyes are gazing toward the examiner's right, otherwise score 0. Smiling Mouth ( Item 72): Score 1 if figure's mouth is smiling, otherwise score 0. Frowning Mouth ( Item 73): Score 1 if figure's mouth is frowning, otherwise score 0. Straight Line Or Slash Mouth ( Item 74): Score I if figure's mouth is a straight line or slash, otherwise score 0. Mouth With Teeth Showing ( Item 75): Score 1 if teeth are present In figure's mojth, otherwise score 0. Object In Mouth (Item 76): Sore I if a i object is presait (i.e., cigar, pipe) in fiÿjre's mouth, otherwise score 0. ReechinoArms ( Item 77); Score 1 If both of figure's a^ms are reaching above the figure's head, otherwise score 0. Outstretched Arms ( Item 78): Score 1 if both of figure's arms are outstretched to side of figure ( approximately perpendicular to the midline of the figure), otherwise score 0. 103 HanoInoArms ( Item 79): Score I lf both of figure's arms are hanging at side of figure with visible space between torso of figure and arms, otherwise score 0. Arms Pressed To Torso ( Item 80): Score 1 1f both of figure's arms are hanging at side of figure with no visible space between torso of figure and arms, otherwise score 0. Arms Inconsistently Positioned ( Item 81 ): Score 1 if each of figure's arms is in e different category among those described above (numbers 77 through 80). Otherwise score 0. Hands Cut-Off ( Item 82): Score 1 if ends of arms are visible wiUi neither hands nor fingers present Otherwise score 0. (Hands hidden behind back of figure or in pockets not scored). Hands Hidden ( Item 83): Score 1 if hands are hiddai behind back of figure or in pockets. Otherwise score 0. Fists (Item 84): Score 1 if hands are made into fists. Otherwise score 0. Talon-Like Fingers ( Item 85): Score 1 If one or more fingers are clearly pointed or knlfe- llke. Otherwise score 0. Agressive Symbols ( I tan 86): Score 1 tor the presence of one or more aggressive symbols (such as guns, knives, clubs, or other symbols of aggression). Otherwise score 0. Oblect(s) Attached To Fioure ( item 87): Score 1 for (he presence of one or more objects attached to, or being held by, the figure ( I.e., tendbag, briefcase, backpack, baseball bat, etc., excluding aggressive symbols). Otherwise score 0. Badcqround Filled-In ( Item 88): Score 1 for the presence of anything drawn in addition to the human figure which Is not attached to or being held by the figure ( i.e., animal, artomobile, building, tree, sun, moon, clouds, raindrops, etc.). Othw^ise score 0. 10 4 Broken Lines ( Item 89): Score 1 if the figure is drawn with broken lines. Otherwise score 0. Buttons ( Item 90): Score 1 for the presaice of buttons (a row of at least three buttons on shirt or pants). Otherwise score 0. Non-Human Or Monster ( Item 91 ): Score 1 if the figure is (^awn as a non-humsn or monster. Otherwise score 0. Multiple Figures On A Sinole Paoe ( Item 92): Score i for the presence of more than one complete figure on the page. Otherwise score 0. Nude Figure ( Item 93): Score 1 for a fully or partially unclothed figure. Otherwise score 0. (Bare feet, short pants, short sleeve shirt are ngt scored). Soldier Or Cowbov/Cowoirl ( Item 94): Score 1 for a fi^jre drawn as a soldier, cowboy/cowgirl. Otherwise score 0. Appendix C FINAL SCORING MANUAL 1 ) Short Fiojre ( Height): Score 1 if figure height is less thaï the criteria listed below, otherwise score 0. Figure height equals the distance between the uppermost point of the figure and the lowermost point of the figure. Articles of clothing such as hats or siioes are Included in the measurement» while other objects ( i.e., handbag, briefcase, backpack, baseball bat, etc.) are not included in the measuranent. Man Drawing - less than 64 mm Woman Drawing - less than 65 mm Self Drawing - less thoi 60 mm 2) Small Figure (Area): Score 1 if fl^ re area Is less than the criteria listed below, otherwise score 0. Figure area equals figure heiÿit multiplied by figure width. Figure width equals the distance betwem extreme left aid right (Mints of fi«jre. As with figure height, articles of clothing are included in the measurement of figure width while other objects are rwt included in the measurement. Man Drawing - less than 2,200 sq mm Woman Drawing - less than 2,450 sq mm Self Drawing - less than 1,600 sq mm 3) Left Placement: Score 1 if txrizontal figure placement is less than the criteria listed below, otherwise score 0. Horizontal figure placement e t^ ls ((FW/2) + ( LD)) / PW, where FW = fiÿjre width, LD = distance from leftmost point of figure to left edge of page, aid PW = the width of the page ( 2 ! 6 mm ). » All measurements are to be made in millimeters. 105 106 Item 3 (continued) Man Drawing - less than 0.33 Woman Drawing - less than 0.35 Self Drawing - less than 0.38 4) Runs-Off Psæ: S ore I if any portion of the figure touches any edge of the paper or runs-off the page, otherwise score 0. 5) Leo Length Uneven: Score 1 if the number resulting from dividing the length of the shorter leg by the length of the longer leg is less than the crita'ia listed below, otherwise score 0. See Figure Ten, below, for measurement of leg length. Man Drawing - less than 0.86 Woman Drawing - less thaï 0.88 Self Drawing - less thoi 0.87 Leg Length Leg Length Leg Length Figure 10 Leo Length Measurement Leg length equals distance between point where leg attaches to body (i e.. crotch) and the firtherm ost point of leg, hebl of foot or shoe, or toe. If leg is bent at the knee, leg length equals distance between point where leg attaches to body and knee plus distance between knee and furthermost point of leg, heel of foot or shoe, or toe. 107 6) Foot lenoth Uneven: Score 1 1f the number resulting from dividing the length of the shorter foot by the length of the longer foot is less than the criteria listed below, otherwise score 0. See Figure Eleven, below, for meesurement of foot length. Man Drawing - less than 0.71 Woman Drawing - less tto i 0.75 Self Drawing - less than 0.7 Foot Length Figure 11 Foot Length Measurement Foot length equals distance between rear of heel «id end of toe or toe-end of shoe. If heel is not shown foot length equals distntce between ankle and end of toe or toe-end of shoe. 7) Torso/Stance Disproportion : Score I if the number rœulting from dividing the width of the narrowest portiez of the torso by the stance width exceeds the limits listed below, otherwise score 0. See Figures Twelve and Thirteen, below, for measurement of tte width of the narrowest portion of the torso and the stance width. Man Drawing - less than 0.25 or greater thwi 1.00 Woman Drawing - less than 0.25 or greater thai 0.96 Self Drawirg - less than 0.27 or greater than 0.90 10 8 Item 7 (continuai) Vidth of Narrovest Portion of Torso Figure 12 Width of Narrowest Portion of Torso Meesurement This measurement is t^dcen at the narrowest portion of the the torso (below neck, if present, and above crotch or point where legs attach). ^ I I—) Stance Vidth Figure 13 Stance Width Meesurement stance width equals the distance between left-most point of foot (or bottom-most portion of leg If foot not present) and riÿ it most point of foot (or bottom-most portion of leg If foot not present). 8) Baseline Drawn: Score 1 (per man, woman, arKl self drawing) if a baseline is drawn ( i.e., yourwl, grass) for figure to stand on. Otherwise score 0. 109 9) Figure Wldth/Heldit Disproportion: Score 1 1f the number resulting from dividing the figure width by the figure height exceeds the limits listed below, otherwise score 0. Man Drawing - less than 0.35 or greater than 0.61 Womai Drawing - less than 0.33 or greater than 0.88 Self Drawing - less tl«n 0.33 or greater than 0.88 10) Head Width/Heioht Disoroportion: Score 1 if the number resulting from dividing the head width by the head height exceeds the limits listed below, otherwise score 0. See Figures Fourteen and Fifteen, below, for measurement of the width and height of head. Man Drawing - less than 0.63 or greater than 1.23 Woman Drawing - less than 0.60 or greater than 1.38 Self Drawing - less than 0.62 or greater than 1.33 Head Width Fiÿjre H Head Width Measurement Head width equals the distance between extreme left and right points of head, following the contour of the face. Measurement does not include ears, hair, or hat. n o item 10 (continued) Head Height r Figurais Head Haiÿit Measurement Head heiÿit equals the distance between the lowermost point of head and the uppermost point of head (including hair or beard but not neck). IF figure is wearing a hat or a piled-high hair style, measure to the top of hat or hair unless the actual top of head is clearly visible. If top of head is clearly visible, measure only to that point. 11 ) Arm Length/Leg Laioth Disorooortion: Score 1 if the number resulting from dividing the average arm length by the average leg length exceeds the limits listed below, o th a^iæ score 0. Average arm length equals üie sum of both arm lengths divided by two. Average leg Imgth ecpjals the sum of both leg Imgths divided by two. See Figire Sixteen for measurement of Arm L e i^ . See F igure Ten for measurement of Leg Length. Mai Drawing - less tiran 0.47 or greater than 1.34 Woiriffli Drawing - less tl%n 0.47 or greater than 2.02 Self Drawing - less than 0.50 or greater than 1.31 n i Item 11 (cOTtInued) Arm Length Figure 16 Arm Length Measurement Arm length equals distance itetween point where arm attaches to the body (from m^n pit) and furthermost point of win. hmd. or finger. If arm is bent at the elbow, arm length equals distance between point where arm sttaches to body and elbow plus distance belweei, elbow and furthermost point of arm. hand, or finger. 12) Arm Lenoth/Fiqure Height Disorooortion: Score 1 if the number resulting from dividing the average arm length by the figure height exceeds the limits listed below, otherwise sctreO. Man Drawirq- less than 0.15 or greater thm 0.38 Wonan Drawing - less thaï 0.13 or greater than 0.37 Self Drawing - less than 0.16 or greater than 0.41 13) Leg Length/Figure Height Disorooortion: Score I if the number resulting from dividing the average leg length by the figure height exceeds tte limits listed below, otherwise score 0. Mm Drawing - less than 0.19 or greater than 0.47 Womm Drawing - less than 0.12 or yeater than 0.43 Self Drawing - less than 0.22 or g ^ te r thwi 0.48 112 14) Hand Area/Flaire Area Disproportion: Score 1 If the number resulting from dividing the average hand area by the figure area is greater than the criteria listed below, otherwise score 0. Hand area equals Hand Height multiplied by Hand Width. See Figures Seventeen and Eighteen for measurement of Hand Height and Hand Width. Average hand y e a equals the sum of both hand areas divided by two. Man Drawing - less than 0.003 or greater than 0.0225 Woman Drawing - less than 0.0025 or greater than 0.0210 Self Drawing - less than 0.003 y yeater than 0.0255 Hand Height Figure 17 Hand Height Measurement Hand height equals distance between uppermost point of hand (including fingers) and lowermost point of hand (including fingers) without regard to orientation of hand. Hend'Vidth Figure 18 Hand Width Measurement Hand width equals distance between extreme left md right points of hand (including fingers) without regard to orientation of hand. 11 3 15) Paœ Rotated: Score 1 (per man, woman, and self drawing) If figure Is drawn with tt» longest dimension of page (279 mm) at ü«top/tîottom. Otherwise score 0. 16) Failed Integration of Bod/ Parts: Score 1( per man, woman, and self drawing) for failure of any portion of the following: Head attached to neck or top of torso, two arms (one if in profile) attached to the top half of the twso (above the half-way mark in the vertical measurement of the torso), and two legs (one if in profile) attached at the bottom of the torso. Otherwise score 0. 17) Hair Omitted: Score 1 (per man, woman, and self drawing) if the figure has rw hair (on head), otherwise score 0. Any attempted representation of hair (on head) cannot be scored as an omission. 18) htose Omitted: Score 1 ( ptr man, woman, md self (Rawing) if the figure's nose is absent, otherwise score 0. Any attempted representation of a nose cwnot be scored as w» omission. 19) Mouth Omitted: Score 1 ((œr man, woman, aid self (Rawing) if the figure's mouth is absent, otherwise score 0. Any attempted representation of a mouth cannot be scored as an omission. 20) Arms Omitted: Score 1 ( p » 'mai. woman, and self drawing) if the figure has no arms, otherwise score 0. Any attempted représentation of arms ( including only one arm ) cannot be scored as an omission. 21) Fingers Omitted: Score 1 (per man, womai, and self drawing) if the figure has no fingers, otherwise score 0. Any attonpted representation of fingers cannot be scored as oi omission. 22) Feet knitted: Score 1 (per mai, woman, and self drawing) if tte figure has r» fæt, otherwise score 0. Any attempted representation of feet ( including only one foot) cginot be scored as an omission. 14 23) Clothino Omitted: Score 1 ( per man, woman, and self drawing) If the figure has no indication of clothing, otherwise so re 0. Any attempted representation of clothes cannot be scored as an omission. 24) Feet Shading Score 1 ( per mai, woman, and self (Rawing) if pencil strokes are present on the figure's feet (foot) which are desigrwJ to fill in æ as in coloring or darkening. Otherwise score 0. 25) Outside Of FiQure Shadina Score 1 ( per man, woman, and self drawing) if pencil strokes are present outside of the fi^jre which are desiyied to fill in an a ^ as in coloring or darkening. Otherwise score 0. 26) Frowning florth: Score 1 (per mai, woman, are! self drawing) if figure's mouth !s frowning, otherwise score 0. 27) Reaching Arms: Score 1 ( per man, woman, and self drawing) if both of figure's arms are reaching above the figure's head, otherwise score 0. 28) Hands Cut-Off: Score 1 ( per mai, woman, and self drawing) if ends of arms a ^ visible with neither hands nor fingers present. Otherwise score 0. ( Hands hidden behind back of figure or In pockets not scored). 29) Adtressive Symbols: Score 1 (per ma», womai, aid self drawing) for the presence of one or more aggressive symbols (such as guns, knives, clubs, or other symbols of aggression). Otherwise score 0. 30) Obiect(s) Attached To Figure: Score 1 (per man, woman, and self drawing) for the presence of aie or more (Ajects attached to, or being held by, the figure ( i.e., handbag, briefcase, backpack, baseball bat, etc., excluding aggressive symbols). Otherwise score 0. 11 5 31 ) Badcorourd F111ed-ln: Score 1 (per man, woman, andssif arawtng) for the presenceof anything drawn in addition to the humm figure whi(A is not attached to or being held by the figure and is not an agyessive symbol (i.e., animal, automobile, building, tree, sun, moon, clouds, raindrops, etc.). Otherwise score 0. Appendix D SLANTED FIGURE SCORING TEMPLATE a s 116 REFERENCES Adler, P. ( 1970). Evaluation of the figure drawing technique: Reliability, Tectorial structure, and diagnostic usefulness. Journal of Consultino and Clinical Psvcholoo/. 35. 52-57. Anastasi.A. (1988). Psvcholooical Testing (6th ed.). New York: MacMillan. Ansbacher,H. L. ( 1952). TheGoodenoughDraw-a-Man test and primary mental abilities. Journal of Consulting Psvcholoov. 1 6 , 176-180. Apfeldorf, M., Randolph, J. J ., & Whitman, 6. L. ( 1963). Figure drawing correlates of furlKjgh utilization in an aged institutionalized population. Journal of Proiective Techniques. 2 7 .158-165. Apfeldorf, M.,& Smith, W. (1966). The representation of the bod/self in human figure drawings. Journal of Proiective Techniques. 50.283-289. Apfeldorf, M., Smith, W., Peixotto, H. E., & Hunley, P. J. (1974). Is the representation of the objective body image in figure (Swings related to personality characteristics of the drawer? Psvcholooical Reports. 3 4 .1015-1020 Arkell,R. N. (1976). Naive prediction of patholoçv from humai figure cb'awings. Journal of school Psvcholoov. 14.114-117. Armstrong, R. 6. & Hauck, P. A. (1961). Sexual identification and the first fi^ re drawn. Journal of Consultino Psvcholotv. 2S, 51-54. Bailey, W. J ., Shinedling, M. M., & Payne, I. R. (1970). Obese individuals'perception of body image. Perceptual and Motor Skills. 31.617-618. 117 11 8 Bakare, C. 6. M. (1972). Social-class differences In the performance of Nigerian children on the Draw-a-Man test. In L. J. Cronbach & P. J. D. Orenth ( Eds.), Mental tests and cultural adaptation ( d p . 355-363). The Hague; Mouton. Baldwin, I. T. (1964). The head-body ratio in human figure drawings of schizophrenic and normal adults. Journal of Proiective Techniques. 28.393-396. Beliak, L. (1950). On the problems of the concept of projection. In L. E. Abt, & L. Beliak ( Eds ). Proiective psvcholoov: Clinical aooroaches to the total person ( o p . 7-32). New York: Knopf. Berdie, R. F. ( 1945). Measurement of adult intelligence by (frawings. Journal of Clinical Psvcholog/. i , 288-295. Bieliauskas, V. J. ( 1960). Sexual identification in children's drawings of human figure. Journal of Clinical Psvcholoo/. 16.42-44. Bieliauskas, V. J ., & Bristow, R. B. ( 1959). The effect of formal » 't training on the quantitative scores of the H-T-P. Journal of Clinical Psvcholoov. 15.5 7 -5 9 Brill, M. (1935), The reliability of the Goodenough Draw-a-Mm test and the validity of an abbreviated scoring method. Journal of Educational Psvcholoov. 2 6 ,7 0 1 - 708. Brown, W. R., & McOiIre, J. M. (1976). Current psyctologlcal assessment practices. Professional Psychology. 7,475-484. Buck, J. N. ( 1948). The H-T-P technique; a qualitative and quantitative scoring manual. Journal of Clinical Psvcholoov. 4 ,3 1 7 - 396. Buros.O. K. (Ed.). (1978). The eioth mental measurement yearbook (Vol. 1). Highland Park,NJ: Gryphon. Burt, C. (1921). Mental and scholastic tests. London: P. S. King and Son. 19 Burton, A., & 5)oberg, B., Ur. (1964). The diagnostic wltdlty of human figure drawings In schizophrenia. Journal of Psvcholog/. 3-18. Carlson, K., Quinlan, D. & Tucker, G., & Harrow, M. (1973). Bod/disturbance and sexual elaboration factors in figure drawings of schizophrenic patients. Journal of Personality L 37,56-63. Carîæn. K.,Tucker,0., Harrow,M.,&Quinian,D. (1971). Bod/image and mental illness. In I. Jakab (Ed.), Psychiatry and ar t M 1- Conscious and unconscious expressive art (pp. 162-169). Basel: Karger. C8ssel,R. H., Johnson, A., & Burns, W.H. (1958). Examiner, ego defense, and H-T-P Test Journal of Clinical Psychology. 14.157-160. Cattell, R. B., Eber, H. W., & Tatsuoka, M. M. (î970). Handbook for the Sixteen Personality Factor Questionnaire. Champa#, It: Institute for Persorality and Ability Testing. Cauthen, N. R , Sandman, C. A., Kilpatrick, D. 0., & Deabler, H. L. (1969). DAP correlates of Sc scores on the MMPI. Journal of Proiective Technioues and Personality Assessment. 33,262-264. Cooke, E. (1885). Art teaching and child nature. London Journal of Education. Cressen, R. ( 1975). Artistic quality of drawings and judges' evaluation of the DAP. Journal of Personality Assessment. 3 9 .132-137. Cundick, B. P. (1970). Measures of intelligence on Southwest Indian students. Journal of Social Psvcholoay. 81.151-156. Dieffenbach, E. W. ( 1977). Koppitz' Human Figure Drawing (HFD) Test: The reliability and clinical validity of its emotional indicators. ( Doctoral dissertation. Southern Illinois University at Carbondale, 1977). Dissertation Abstracts International. 3 8 .6055A 120 Dunn,J. (1967). Inter-andlntre-rater reliability of the new Ooodenoigh-Harrls Draw-A-Man Test Pa^ceotual and Motor Skills. 2 4 .269-270. EbeI.R. L. (1961). Must all test be valid? American Psvcholoalst. 16.640 -647. Ellis, R. (1953). Comoarlson of scores on OoodBRaah Drw-A-Man. Revised Stanford-Binet. md Wachsler Ir.tsViioBncB Scale for Children, obtained from a group of children seen in a psychiatrie clinic. Unpublished master's colloquium paper, University of Minnesota Exner.J. E.,Jr. (1962). A comparison of the human figure drawings of psychoneurotics, character disturbances, normals, and subjects experiencing experimentally-induced fear. Journal of Proiective Techniques. 26,392-397. Falk, J. D. (1981). Understanding children’s art; An analysis of the literature. Journal of Personality Assessment. iS , 465-472. Fay, H. M. (1923). Une méthode pour le dépistage des æriérés dans les grandes collectivités d’enfants d’âge scolaire. Bulletin de la Lioue D’Hvoiene Mentale. Feldman, M. J., & Hunt, R. 3. (1958). The relation of difficulty in drawing to ratings of adjustment based on human figure drawings. Journal of Consultino Psychology. 22. 217-219. F isher ,8.(1959). Body reactivity gradteits and figure drawing variables. Journal of Consultino Psychology. 21,54-59. Frank, L. K. (1948). Proiective Methods. Springfield, IL: Char le C. Thomas. Goldstein, A. P., & Rawn, M. L. (1957). The validity of interpretive signs of aggression in the drawing of the human figure. Journal of Clinical Psychology. Ü , 169-171. 121 Goodenough, F. L. ( 1926). Meesurement of intelHoence hv drawings. New York: Harcourt Brace & World. Gollling.S. H. (1985). Comparison of the reliability of the Goodenouoh Draw-A-Man Test with the Naollerl Draw A Person Test: A Qiantltatlve Scoring System. Unpublished master’s thesis, The Ohio State University, Columbus. ^nzburg, H.C. (1955). Scope and limitations of the Goodenough drawing test method in clinical work with mental defectives. Journal of Clinical Psvcholoov. J l , 8-15. Hammer, E. F. (1958). The clinical application of proiective drawings. Springfield, IL: Cho'îœC. Thomas. Hanmer, E. F. (1959). Critlcpje of Swensen's "Empirical evaluations of humoi figure drawings." Journal of Proiective Technioues. 2 1 , 30-32. Hammer, E. P. (1969). DAP: Back against the wall? Journal of Consultino aid Clinical Psychology. 33.151-156. Hammer, E. F., & Piotrowsky, Z. A. (1953). Hostility as a factor in the clinician's personality as it affects his Interprétation of prqjectlye drawings. Journal of Protective Techniques. 1 2 ,2 1 0 -2 î 6. Hanmer, M., & Kaplan, A. M. (1964 a). The reliability of profile ami front-facing directioTS In childrei's drawings. Child Development. 35,973-977. Hammer, M.,& Kaplan, A. M. (1964 b). The reliability of sex of first figure drawn by children. Journal of Clinical Psychology. 2 0 .251 - 252. Hammer, M.,& aid Kaplei, A. M. (1964 c). The reliAility of size of children's drawings. Journal of Clinical Psychology. 20.121-122. 122 Hammer, M.,& Kaplan, A.M. (1966). The rellabrily of children’s human figure drawings. Journal of Clinical Psychology. 22.316-319. Handler, L.,&Reyher,J. (1964). The effecteof stress on the Oraw-A-Person Test Journal of Consulting Psvcholocv. 28.259-264. Hanvik, L. J. (1953). The Goodenough test as a measure of intelligance In child psychiatric patients. Journal of Clinical Psvcholoo/. ^ 71 -72. Harris. D. B. (1963). Children's drawings as measures of intellectual maturity. New York: Harcourt Brace & World. Harris, D. B.,& Roberts, J. ( 1972). Intellectual maturity of children: Demooraohlcand SQciometricfactcrs ( DHEW, Vital and Health Statistics Series I I , No. 116). Hathawy, S. R.,& McKinley, J.C. (1943). The Minnesota Multlohasic Personality Inventory (rev. ed.). Minneapolis: University of Minnesota Press. Haworth, M., & Normington, C. (1961). A sexual differentiation scale for the D-A-P Test Journal of Proiective Techniques. 25,441-450. Heinrich, P., & Triebe, J. K. (1972). Sex p re fa m e s in children’s human figure drawings. Journal of Personality Assessment. 3&, 263-267. Hiler, E. W., & Nesvig, D. ( 1965). An evaluation of criteria used by clinicians to infer pathology from figure drawings. Journal of Consulting Psvcholocv. 22.520-529. Howell, K.W.,Zudcer, S. H.,&Moreheed,M. K. (1985). Multilevel Academic Survey Test. Columbus, OH: Charles E. Merrill. Huberty.C. J. (1984). issues in the use and interpretation of discriminant oialysis. Psvcholooical Bulletin. 9 5 . 156-171. 1 2 3 Hunt, R. 6., & Feîdmen, M. J. (1960). Bal/ Image and ratings of atflustment on human figure drawings. Journal of Clinical Psychology. 16.35-38. Joizen, W. B., & Oœ, W.C. (1975). Clinical and sign prediction; The Draw-A-Person and female homosexuality. Journal of Clinical Psvcholoov. 3 1 .757-765. Jegede, R. 0., & Bamboye, E. A. (1981). Self-concepts of young Nigerian adolescents. Psychological Reports. 4 9 .451 -454. Johnson, FA. (1 972). Figure drawings in subjects recovering from poliomyelitis. Psychosomatic Medicine. 34. 19-29. Jones, L. W., & Thomas, C. B. (1961). Studies on figure drawings: A review of the literature (1949-1959), Psychiatric OuarterlvSuPDlement. 35.212-261. KtfiHI.S. (1984). Human figure drawings in adults: An update of the empirical evidence, 1967-1982. Canadian Psychology. 25,269-292. Kahn, M.W.,& Jones, N. F. (1965). Human figure drawings as predictors of admission to a psychiatric hospital. Journal of Projective Techniques. 2 9 .319-322. Kamano, D. K. (1960). An investigation of the meaning of liuman figjre drawing Journal of Clinical Psychology. 16,429-430. Katzaroff, M. D. (1910). Qu"8t-ce que les enfants dessinait? (What do children draw?) Archives de Psychologie. 9.125. Kerschensteiner, D. G. (1905). Die Entwickelunqder zeichnerischen Beoabuna Munich: Gerber. Klepsch, M., & Logie, L. ( 1982). Children draw «id tell. New York: Bruner/Mazel. 124 Kkpfer, W. 0., & Taulbee, E S. (1976). Pn^ecttve tests. Annual Review of Psvchotow. 27. 543-567. Koppitz, E. M. (1968). Psychological evaluation of children's human figure drawlnœ. New York: Grune& Stratton. Koppitz, E. M. (Î9S4). Psychological evaluation of human figure drawings t»v middle school Duoils. New York: Grune & Stratton. Kuhlman, T. L.& Bieliauskas. V. J. (1976). A comparison of black aid white adolescents on theHTP. Journal of Clinical Psvcholoo/. 3 2 .728-731. Lair, C.Y„&Trapp, E. P. (1960). Performance decrement on the H-T-P Test as a function of adjustment level. Journal of Clinical Psvcholoo/. 431. Laosa, L. M.,Swo-tz, J. D , & Diaz-Ouerrero, R. (1974). Perc^tiial-cognitiveaid personality development of Mexican and Ar^lo-Amo'icai children as measured by human figure drawings. Develoomental Psvcholoov. 10.131-139. Lapkin, B., Hillaby, T. 0., & Silverman, L. (1968). Manifestations of the schizophrenic process in figure drawings of adolescents. Archives of Oeneral Psvchiatrv. J â , 465-468. Leichtman, S. R., Burnett, J. W., & Robinson, H. M., Jr. (1981). Body image concerns of psoriasis patients as reflected in human figure drawings. Journal of Personality Assessment. 45.476-484. Leland, H. (1983). Diagnosis of central nervous system disorders with prtqective techniques. In C. J. Golden & P. J. Vincente (Eds.). Foundations of clinical neuroosvcholoav(DD. 143-161). New York: Plenum. Lembke.W. (1930). Uber &1chnungen von "frechen" und "schuchternai" Schulkindern. (The significance of (kawings by "shy" and “bold" pupils.) Z. f. padPavchol.. (Oct) 459-469. (Fran Psvctolooical Abstracts. 1930, §, Abstract No. 1657) 125 Levinstein,S. (1905). Klnderzelchnunoen tiiszum 14lebensiahr. Leipzig: R. Volÿlander Yerlag. Levy, I. (1971). The Harris-Qoodenough drawing test end educeble mentally retarded adolescents. Jourrel of Mental Deficiencies. 75,760-761. Lewinsotm,P.M. (1965). Psychological cu r'elatescf overall quality of figure drawings. Journal of Oonsultino Psvcholoov. 29,504-512. Litt,S.,&Margoshes,A. (1966). Sex-change in successive Draw-a-Person Tests. Journal of Clinical Psychology. 22,470. Loney.J. (1971). The sun as a measure of dependency in chiltfren's drawings. Journal of ClinJcal Psvchotav. 27.513-514. Lorge, I., Tuckman, J., & Dunn, M. B. (1958). Human figure dw ings t)y younger and older adults. Journal of Clinical Psvcholoov. 14.56-58. LowenfeW.V. (1947). Creative and Mental Growth ( 2nd ed ). New York: MacMillan. Lowenfeld, Y. (1952). Creative and Mental Growth ( 3rd e l). New York: MacMillan. Lubin, B., Larsen, R.M.,&Mato'az 20,J.D. (1984). Psychological test usage in the United States: 1935-1982. American Psvcholooist. 3 9 ,451-454. Lubin, B., Wallis, R. R.,& Paine, C. (1971). Patterns of psychological test usage in the United States: 1935-1969. Professional Psvcholotv. 2.70-74. Ludwig, D. J. (1969). Self-perception and the Draw-A-Person Test. Journal of Proiective Technioues end Persnnalitv Assessmept 2 5 7 -2 6 1. Luquel, 0. H. (1913). Les Dessins d’un Enfant etude psvcholoaioue. (The (rowings of a child.) Paris: Librairie Felix Alcan. 125 Luquet, G. H. (1927). Le réalisme intellectisl dans l'art primitif: I. Figuration (te l'invisible. ( Intellectual realism in primitive ar t I. Representation of the invisible.) Jornal de PsicholoQia. 2 4 .765- 797. LiK]uet, G. H. (1929). L'evolution du dessin enfantin, (The development of (Aildlsh drawing.) Bulletin Soc Binet. 29.145-163. Mechover.K. (1949). Personality oroletAlon in the drawing of the human figure. Springfield: Charles 0. Thomas. Maloney, M. P.,& Glasser, A. ( 1982). An evaluation of the clinical utility of thd)raw-A- Person Test Journal of Clinical Psvcholoov. 38.183-190. Manganyi, N. C. (1972). Body image bounctey differentiation and self-steering behaviour in African paraplegics. Journal of Personality Assessment. 45-49. M ^iss, H.C.,& Strumpfæ, D. J. W. (1965). DAP body-image disturbance scale and (juality of drawing. Perceptual and Motor Skills. 21.196. Marzolf, S. S., & Kirchner, (1972). Characteristics of House-Tree-Person drawings of college men and women. Journal of Proiective Techniques and Personality Assessment. 138-145. McCarthy, D A. (1944). A study of the reliability of the Goodenough drawing test of intelligence. Journal of Psvcholiw. J 8 , 201 -206. McCarthy, D. A. (1972). Manual for the McCarthy Scales of Children's Abilities. San Antonio; Psychological Corporation. McCurdy, H. G. (1947). Group and individual variability on the Goodenouÿi Draw-a-mm test Journal of E(kcationa1 Psvchol(w. 428-436. ticElwee, E. w. (1932). The relladiltîy of the 0Godsnougli1ntBl1lgaK3 test usod with sub-norrael children fourteen years of age. Journal of Applied Psychology. 16.217- 218. McHugh, A. F. (1963). H-T-P proportion and perspective In Negro, Puerto Rican, and White children. Journal of Clinical Psvchokw. 19.312-313. McLachlan, J. F. C., & Head, Y. B. (1974). An impairment rating scale for human figure drawings. Journal of Clinical Psvcholoov. 30.405-407. Meehl ,P. E.,& Rosen, A. ( 1955). Antecedent probability and the efficiency of psychometric signs, patterns, or cutting scores. Psychological Bulletin. 52.194-216. Mogar, R. E. (1962). Anxiety indices in human figure drawings; A replication and extension. Journal of Consultino Psychology. 26.108. Morris, W. W. (1955). Ontogenetic ctenges in adolescence reflected by the drawing-human- figures technique. American Journal of Orthoosvchiatrv. 2 5 ., 720-728. Mortensen, K. V. (1984). Children's human figure drawings: Development, æx differences, and relation to psychological theories (Vol. 1 ). Dansk: Psykologisk Forlag. Murray, D.C., & Deabler, H. L. (1958). Drawings, diagnoses, and the clinician's learning curve. Journal of Proiective Techniques. 22.415-420. Nag]feri,J.A. (1985). Matrix Analogies Test - Short Form. New York: Psychological Corporation. Nag1ieri,J. A. (1988). Draw A Person: A ajantitative sccrino system. New York: Psychological Corporation. Nichols, R., & Slrumpfer, D. J. W. (1962). A factor analysis of Draw-A-Person Test scores. Journal of Consulting Psvcholotr/. 26.156-161. Î2 8 Ochs, E. ( 1950). Changes in Qoocfaiough drawings essecMed with rt«ngBS In social atilüstmenL Journal efClinical Psvchotay. 6.282-284. Ogden, D. P. (1977). Psvdwhaonostlcs end personality assessment- A handbook (2nd ed). Los Angeles: Western Psychological Services. Ohio Department of Education. (1982). Rules for the education of handicsooed diildren. Worthington, OH: Author. Ottenbacher, K. (1981). An investigation of self-concept and bod/ image in the mentally retarded. Journal of Clinical Psychology. 37.415-418. Po'tridge, L. ( 1902). Oilldren's drawings of men and women. In E. Barnes (Ed), Studies in education (Vol. 2),(pp. 163-179). PedToux, R , KcAler, M , & Oirard, Y. (1947). Réflexions sur l'évaluation de 1 "intelligence chez les enfants Irréguliers. (Reflections on the evaluation of intelligence in abnormal children). Journal Méd Lvon. 28.337-343. ( From Psychological Abstracts. 1947, 21, Abstract No. 4370). Phillips, W. M , & Phillips, A. M. (1976). Similarity between complexity on role construct rdJerlory technique and articulation on Draw-A-Person test for patients md nm-patients. Perceptual and Motor Skills. 4 3 .1256-1258. Pihl,R.O.,& Nimrod, 6. (1976). Tte reliability and validity of the Draw-A-Person test in 1.0. Old personality assessment. Journal of Clinical Psvchoiooy. 52,470-472. Pikulski, J. (1972). A comparison of figure drawings and WISC IQs among disabled readers. Journal of Learning Disabilities. 5.156-159. Pryor, F. A. & Butler, J. R. (1969). Test of the unimportance of particular item content using the H-T-P «id the MMPI. Psychological Reports. 2 4 .989-990. 129 Rstîn.A. !. (E l). (1981). Assessment with prriætlvs technlûüs. Ne* York: %r1ngcT. Rabin,A. l.,&Limuaco,J. (1959). Sexuel differentiation and Filipinochildren as reflected in the Draw-A-Persori Test Journal of Social Psvcholoo/. 5 0 .20 7 -2 1 1. Reichenberg-Hacicelt, W. (1953). Changes in Ooodenough drawings after a gratifying experience. American Journal of Orthoosvchiatrv. 23.502-517. Rey.A. (1946). Epreuves de dessin témoins du développment mental. I. Archives Psychologie Geneve, ü . 369-380. Ricci, C. (1887). L'Artedei Bambini. Bologna: Zanichelli editore. Roback, H. B. (1968). Human figure drawings; Their utility in the clinical psychologist’s armanenterium for personality assessment Psvcholooical Bulletin. 7 0 .1-19. Rorsd%ch,H. (1942). Psychodiagnostics: A diagnostic test based on perception ( P. Lemkau & B. Kronenberg, Trans.). Berne: Huber ( Ist Germai e l published 1921; U.S. distributor, Grune & Stratton). Rosenberg, LA. (1965). R^ld changes In overt behavior reflected in the Draw-A-Person: A case report Journal of Proiective Techniques. 29.348-351. Rosenthal, M.M.&Beutel, N.J. (1981). Movement and body-image: A preliminay study. Perceptual aid Motor Skills. 53,758. Rotter, J. B. ( 1966). Generalized expectancies for internal versus external control of reintorcement Psychological Monographs. 80. (Whole No. 609). Rottersman, L. A. (1950). A canoarison of the IQ scores on the new revised Staiford-Binet. Form L.theWechsler Intelligence Scale for Children, and the Goodenouoh “Draw-a-man* test at the six veer age level. Unpublished master’s thesis. University of Nebraska. 130 Rubin, J. A., Raglns, N., SctMchter, J., & Wimberly, F. ( 1979). Drawings by schizophrenic and non-schizophrenic mothers and their lAildren. Art Psvtfwtheraov. 6 ,1 63-175. Russakoff, L. M., Fontana, A F., Dowds, B. N.,&Harris, N. ( 1976). Psychological differentiation and psychotherapy. Journal of Nervous m i Mental Disease. 163. 329-333. Settler, J. M. (1982). Assessment of children's intelligence snA special abilities (2nd ed.). Boston; Allyn & Bacon. Sattler, J. M. (1988). Assessment of children ( 3rd ed.). San Diego: Jerome M. Settler. Schaefer, C. E., & Sternfleld, M. (1971). Comparative validity of the Harris Quality and Point Scale. Perceptual and Motor Skills. J i , 997-998. Schaeffer, R. W. (1964). Clinical psychologists'ability to use the Draw-A-Person Test as an Indicator of personality adjustment Journal of Consulting Psvcholoov. 28.385. Sdimidt, L. D. & McGowan, J. F. (1959). The differentiation of human figure drawings. Journal of Consulting Psychology. 23.126-133. Scott, LH. (1981). Measuring intelligence with the Goodenough-Harris Drawing Test Psvcholooical Bulletin. 89.483-505. Slm m , J. ( 1962). Intraindividual re^wnse variability in Figure Drawing Tasks. Journal of Proiective Techniques. 26.105-111. Sherman, L. J. (1953). The influence of artistic quality on judgments of patient and non- patient status from human figure (frawlngs. Journal of Proiective Technioues. 2 2 ,3 3 8 - 340. Silverstein, A B. (1966). Anxiety and quality of human figure drawings. American Journal Of Mental Deficiency. 2fi, 607-608. 131 Silverstein, A. B.,& Robinson, H. A. (1961). The representation of physique In children's figure drawings. Journal of Oonsultino Psvcholoov. 25.146-148. Smith, P.O. (1937). What the Godenough Intelligence test irionurss. Psychological Bulletin. 34,760-761. Snyder, R. T., & Gaston, D. S. (1970). The figure (fr'awing of the first grade child: Item analysis and comparison with Koppitz norms. Journal of Clinical Psvcholoov. 2 6 ,3 7 7 - 383. Solar, D.,Bruehi,D., &Kovacs, J. (1970). The Draw-A-Person Test: Social conformity or artistic ability? Journal of Clinical Psvcholoov. 26.524-525. Strieker, G. ( 1967). Actu»*i8l, reive clinical, and sophisticated clinical prediction of pathology fr-om figure drawings. Journal of Consultino Psychology. 31.492-494. Slrumpfer, 0. J. W. (1963). The relation of Draw-A- Person test variables to age and chrofiicity in psychotic groups. Jourrel of Clinical Psvcholow. J l , 2 0 8 -2 1 1. Suinn, R. N., &Oskamp, S. (1969). The predictive validity of proiective measures: A fifteen year evaluative review of research (Chapter 8). Springfield: Charles C. Thomas. Sully, J. (1895). Studies of childhood. London: Longmans, Green. Suntfcerg, N. D. (1961). The practice of psychological testing In clinical services in the United States. American Psvcholooist. 16.79-85. Swensen,C. H. (1957). Empirical evaluations of human figure îk'awir^. Psychological Bulletin. 54.431-466. Swensen, C. H. (1968). Empirical evaluations of human figure drawings: 1957-1966. Psychological Bulletin. 7 0 .20-44. 132 Terman.L. M. (1916). The measurement of Intelligence. Boston: Houghton Mlfflln. Thurstone.L. L (1938). Primary mental abilities. Psychometric Monooraohs. Na 1. Tolor, A., & Tolor, B. ( 1974). Children's figure drawings and changing attitudes towards sex roles. Psvcholooical Reports. 34.343-349. Uçman.P. (1972). A normative stud/of the Ooodanough-Harris Drawing Test on a Turkish sample. In L. J. Cronbach & P. J. D. Drenth (Eds.). Mental tests end cultural adaptation (pp. 365-370). TteHague: Mouton. Urban, W.H. (1963). The Draw-A-Person catalogue for interpretive analysis. Los Angeles: Western Psychological Services. Van Dyne, W. T., & Carskadon, T. G. (1978). Relationships among three components of self- conc^l and sane-sex and (çposite-sex human figure drawings. Journal of Clinical Psvcholoov. 34,537-538. Vane, J. R.,&Eisen, V. W. (1962). TheOoodenouÿi Draw-A-Man Test andsigis of maladjustment in kindergarten children. Journal of Clinical Psychology. J 8 , 276-279. Vane, J., & Kessler, R. T. ( 1964). The Goodenough Draw-a-Man Test Laq term reliability md validity. Journal of Clinical Psvcholocv. 20.487-488. Vine/, L. L.,Aitken,M.,& Floyd, J. (1974). Self-regard and size of humai figure drawing: An interactioral cralysis. Journal of Clinical Psvcholoov. 581-586. Wade,T. C.,&Bdcer, T. B. (1977). Opinions and use of psychological tests: A survey of clinical psychologists. American Psvcholooist. 32.874-882. Walsh, W.B.,& Betz, N. (1985). Tests and assessment. Englewood Cliffs, NJ: Prentice-Hall. 13 3 Wanderer, Z. E. (1969). Validity of clinical Judgements based on human figure drawings. Journal of Consulting and Clinical Psvchoiooy. 33.143-150. Watson, 0.0. (1967). Interjudge agreement of Draw-A-Person diagnostic impressions. Journal of Protective Technioues. 5 1 .42-45. Wechsler.D. (1949). Manual for the Wechsler Intelligence Scale for Children. San Antonia Psychological Corporation. Wechsler.D. (1955). Manual for the Wechsler Adult Intelligence Scale. San Antonia Psychological Corporation. Whitmyre, J. W. (1953). The significance of «*t1st1c excellence In the judgment of atgustment inferred from human figure drawings. Journal of Consultino Psvcholoov. 17, 421-422. Wlldman, R. W..& Wlldman, R. W. II. (1967). The practice of clinical psychology In the United States. Journal of Clinical Psvcholoov. 23.292-296. Williams, J. H. (1935). Validity and relii*ility of the Goodenough Intelligence Test. Schools and Society. 41,653-656. Yater, A., Barclay, A., & MoGlllIgm, R. (1969). Inter-rater reliability of sctring Goodenough-Harrls Drawings by disadvantaged preschool children. Perceptual and Motor Skills. 28.281-282. Y^)sen,L N. (1929). The reliability of the Goodenough drawing test with fedaleminded subjects. Journal of Educational Psvcholoov. 20.448-451. Ziv,A.,&Shechori,H. (1970). Human figure as a measure of social adjustment in school. Journal of School Psychology. 8.152-153. 134 Zuk, 6. H. (1962). Relation of mental age to size of figures on the Draw-A-Person Test Perceptual and Motor Skills. 14.410.