FAMlLY DRAWINGS, INDIVIDUAL] f , GROUP IDENTITY AND GROWTH 1 - 'M$mmwmmmmu‘-< wmmmmmumwmm" mmmmRmmv lwo‘ W LIBRARY a" Michigan State E University . This is to certify that the thesis entitled FAMILY DRAWINGS, INDIVIDUAL/GROUP IDENTITY AND GROWTH presented by Franklin R. Kirby has been accepted towards fulfillment of the requirements for Ph.D. degree in Psychology - - l / /] + Date May 14, 1970 0-169 1 ABSTRACT FAMILY DRAWING, INDIVIDUAL/ GROUP IDENTITY AND GROWTH By "\ Franklin R. Kirby Draw-A-Family pencil drawings were collected from u56 students in two public schools, grades one through six, and from 8“ disturbed students. The scoring system, based on Fromm's theory of high group identity/low individ- ual identity -- low group identity/high individual identity as related to growth and truncation of growth, utilized a Scoring Matrix with High and Low Growth columns, and High and Low Group Identity Rows. The resulting four quadrants were labelled Dependence, Overdependence, Independence and Counterdependence. Two techniques were developed for sort- ing the drawings into quadrants. One scoring system, the Position sort, is objective, being based on the position of the "me" in the family draw- ing. The other, the Relational sort, is subjective, involv- ing Judgments of relations like closeness-distance, and affection-encroachment among family members. Two independent Judges achieved an overall agreement of 74% in this second sorting system. Franklin R. Kirby The hypotheses were that younger normal children would tend to fall in the dependent quadrant of the Scor— ing Matrix, with older normals shifting to independence; that younger desturbed children would cluster in the overdependent quadrant, with older disturbed tending more toward counterdependence; and that, overall, normal draw- ings would cluster into the High Growth column, with dis- turbed drawings falling in the Low Growth column. Results with the Position technique support most of the hypotheses for boys, but not for girls. This difference results from assumptions about the meaning of the "me" drawn next to the parents. In older boys, "me" next to the mother had been classed as dependence, whereas it obviously implies overdependence. Younger girls placing the "me" next to mother were classed as independent, instead of the more obvious dependent; and the reverse was true when "me" was drawn next to father. With these adjustments to the Position scoring system, all the hypotheses were sup- ported, except for discrepancies related to social class and race. Normal lower-class children's drawings tended to fall into the Low Growth column of the Scoring Matrix, expecially for boys, much as did the disturbed children's. Normal black children's drawings disproportionately fell into the counterdepence quadrant. The adjustment of lower class children, in light of their family drawings, could involve less closeness between child and parents at earlier ages. For the blacks, the use of isolation as an adaptive Franklin R. Kirby technique appeared likely. An additional finding was that disturbed children placing the "me" next to the like-sex parent tended to represent that parent as impotent, "goofy" or "gimpy" looking. With the Relational scoring system, hypotheses were generally supported for both sexes, although there were much the same discrepancies for the lower class and blacks as in the Position sort. There was also a similar tendency for the youngest normals to collect in the overdependent quadrant. Usefulness of the sorting techniques might include research, clinical practice and group screening for emo- tional problems, However, the Relational sort seems inap- prOpriate for the latter because of it's complexity. The Position sort is simplistic and objective, and shows prom- ise as a screening device - but only after validation of the modifications derived from.the'data in the study. FAMILY DRAWINGS, INDIVIDUAL/ GROUP IDENTITY AND GROWTH By Franklin R. Kirby A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1970 DEDICATION To Lauri and Ken Jay 11 ACKNOWLEDGMENTS I want to thank some very special people who went out of their way to help on this project. Dr. John Cochran, Kalamazoo School Superintendent, and a true professional, made the study possible, while Carl Czukna's wholehearted investment eased the way. Dr. Charles Hanley's patient skill helped shape the final product. Dr. Bill Kell silently perservered. My wife, Caroline gave unstintingly of all of her talents. Lauri and Ken Jay helped daddy by not playing with all the neat pictures. iii TABLE OF CONTENTS Page ACKNOWLEDGMENTS . . . . . . . . . . . . . . iii LIST OF TABLES . . . . . . . . . . . . . . vi LIST OF FIGURES . . . . . . . . . . . . . . ix Chapter I. PROBLEM . . . . . . . . . . . . . . . 1 Review of the Literature . . . . . 2 A System for Analyzing the Family Drawing: Theoretical and Empirical Foundations . . . 8 II. HYPOTHESES . . . . . . . . . . . . . . 19 III. METHOD . . . . . . . . . . . . . . . 22 Subjects . . . . . . . . . . . 22 Collecting the Drawings . . . . . . . . . 25 Scoring . . . . . . . . . . . . 26 IV. RESULTS . . . . . . . . . . . . . . . 29 Position Analysis . . . . . . . . . . . 30 Additional Findings . . . . . . . . . . A9 V. DISCUSSION . . . . . . . . . . . . . . 57 Position Analysis--The Hypotheses . . . . . 57 Position Sort Scoring . . . . . . 61 Position Analysis—-"Me" Placement . . . . . 63 Relational Sort Analysis . . . . . . . 67 Uses for the Family Drawing and the Sorting Techniques . . . . . . . . . . 75 iv APPENDICES . . . . . . . . . . . . . . . 77 Appendix A . 78 Appendix B . 80 Appendix C . 83 Appendix D . 85 Appendix E . 90 Appendix F . 94 Appendix G . 97 Appendix H . . 100 Appendix I . 106 BIBLIOGRAPHY . . . . . . . . . . . . . . . llO Table 10. ll. 12. LIST OF TABLES Scoring Matrix Observed and Expected Frequencies of Family Drawings in Position Analysis: Grades I—III, Quadrant 1 vs. Quadrants 2,3 and A . . . Position Analysis of Family Drawings: Grades I-III vs. Grades IV—VI, Quadrant 1 vs. Quadrant 2 . . . . . . . . . . Position Analysis of Family Drawings by Sex, Class and Growth . . . . . . . . . Position Analysis of Normal and Disturbed Family Drawings: Age and Growth . . . Position Analysis of Normal Family Drawings: Sex, Age and Growth . . . . . . . . Position Analysis of Normal and Disturbed Family Drawings: Age and Group Identity Position Analysis of Family Drawings for Older Children (Age 10- 1A): Adjustment and Growth . . . . . . . . Relational Analysis of Family Drawings, Grades I-VI: High vs. Low Growth . Relational Analysis of Family Drawings in Grades I-III: Quadrant 1 vs. Quadrants 2, 3 and A . . . . . . . . Relational Analysis of Family Drawings: Younger vs. Older, Quadrant 1 vs. Quadrant 2 . Relational Analysis of Family Drawings and Age: High vs. Low Growth . . vi 0 33 3A 36 37 38 A0 A1 A3 A3 AA A6 Table 13. 1A. 15. l6. 17. 18. 19. 2O. 21. 22. 23. 2A. 25. 26. 27. 28. 29. Page Relational Analysis of Normal Family Drawings and Age: High vs. Low Family Group Identity . . . . . . . . . . . A8 Relational Analysis of Family Drawings, ages lO-lA: Normal vs. Disturbed, High vs. Low Growth . . . . . . . . . . . A9 Family Drawings in Each Quadrant: Position Analysis of Normal Boys and Girls . . . . . 51 Family Drawings in Grade I with "me" Next to Parents: Middle vs. Lower Class . . . . 53 Family Drawings and Placement of "me": Normal vs. Disturbed Boys and Girls . . . . 5A Per cent Placement of "Me" Next to Mother and Father, Normal and Disturbed Boys and Girls 0 e o o o o o o o o o o 63 Observed and Expected (50%) Frequencies of Family Drawings, Normal and Disturbed, Quadrants 3 and A . . . . . . . . . . 7A Normal Sample: Boys and Girls by Age and Grade . . . . . . . . . . . . . 8A Emotionally Disturbed Sample: Boys—Girls, Age, Source . . . . . . . . . . . 8A Normal Position Sort: Grade and Sex . . . . . 86 Normal Position Sort: Age and Sex . . . . . 88 Disturbed Position Sort: Age and Sex . . . . 89 Normal Relational Sort: Grade and Sex . . . . 91 Normal Relational Sort: Age and Sex . . . . . 92 Disturbed Relational Sort: Age and Sex . . . . 93 Distribution of Normal Family Drawings, Position Sort: Grade, Quadrant 1 vs. Quadrants 2, 3 and A . . . . . . . . . 95 Normal and Disturbed Family Drawings, Position Sort: Age 6-9 vs. 10—13, High vs. Low Family Identity . . . . . . . 96 vii Table 30. 31. 32. 33- 3A. 35. 36. 37. 38. 39. A0. A1. A2. A3. AA. Drawings, Relational Sort: Normal and Disturbed by Sex, Class and Race; High vs. Low Growth . . Normal Family Drawings, Boys and Girls, Relational Sort: Younger vs. Older, High vs. Low Growth Normal and Disturbed Family Drawings, Relational Sort. Sex and Age, High vs. Low Family Identity Normal Drawings, Position: Growth, Sex, Class, Race . . . . . . Normal Family Drawings in Quadrants, 25% Expected Frequency, Position Sort Normal Family Drawings in Quadrants 1 and 2, Position Sort: Sex and Grade . . Drawings in Quadrants, Normal vs. Disturbed; Position . . . . . . . . . . . Normal Family Drawings. Placement of "Me," Based on 12-1/2% Chance Expectation; Boys and Girls . . . . . . Normal Family Drawings: Placement of "Me," Boys vs. Girls . . . . . . . . . Placement of "Me" in Family Drawings: Normal vs. Disturbed . . . Normal Family Drawings, Relational Sort: High vs. Low Growth, by Sex, Class and Race . . . . . . . . ' Distribution of Normal Family Drawings by Quadrant, Boys vs. Girls, Relational Sort Distribution Normal Family Drawings by Quadrant, Middle vs. Lower Class, Relational Sort . . . . . Distribution of Normal Drawings by Quadrant, Black vs. Lower Class White, Relational Sort . . . . . . . . . . . . Family Drawings by Quadrant: Normal vs. Disturbed, by Sex, Class and Race; Relational Sort . . . . . . viii Page 98 98 99 101 101 102 103 103 10A 105 107 107 108 109 LIST OF FIGURES Figure Page 1. Per cent of Boys and Girls in Quadrants at Each Grade: Position Analysis . . . . . 31 2. Per cent Normal and Disturbed in Quadrants 1 and 2: Position Analysis . . . . . . . 35 3. Per cent in High Group Identity Rows, Normal and Disturbed: Position Analysis . . . . . 39 A. Per cent of Boys and Girls in Quadrants By Grade: Relational Analysis . . . . . . A2 5. Age and Per cent Normal and Disturbed, High Growth Column of Matrix: Relational Analysis . . . . . . . . . . . . . A5 6. Per cent in High Group Identity Row, Normal and Disturbed: Relational Analysis . . . . A7 7. Per cent of Normal Boys and Girls Drawn Next to Each Parent in Grades I-VI . . . . . . 52 8. Examples of Normal, "Goofy" and "Gimped" Fathers Drawn by 9-year old Boys . . . . . 65 ix CHAPTER I PROBLEM The child's drawing of his family has been used for some time by child guidance workers as an aid in under- standing disturbed children. However, this technique has been used primarily on an "intuitive" basis, because there are no norms which differentiate drawings on the basis of age, sex or emotional well-being. Although other drawing techniques have been widely studied, they do not provide the clues to family dynamics and interactions which are of such moment in the world of the child. The family drawing is readily obtainable from groups of children, and could potentially be of help as a rapid screening device for school personnel concerned with the troubled or incipient problem child. The goal of this study is to examine a number of children's drawings of a family in the light of a theory of group and individual identity, across age, sex, race and adjustment, in an attempt to establish some workable norms for practical use. In essence, this will result in a theoretical, developmental and clinical study. Review of the Literature Although the literature teems with articles on children's drawings (DAP, DAM, HTP, DAA, etc.), there appear to be only nine articles on the family drawing technique. In 1931, a child psychiatrist (Appel, 1931) had his clients draw pictures of themselves and family members separately. Then be elicited comments from the children about their drawings, interpreted them psychoanalytically, and thus felt he gained insight into their problems. How- ever, there were no drawings of the family as a group, and thus no possibility of looking at the family structure and group identity. The first full-fledged use of the family drawing was reported by Wolff (19A6), who was searching for a unifying concept with which to help him understand the behavior of the preschool child. Wolff looked at a number of family drawings of nursery school children and presented indi— vidual cases to elaborate on various qualitative factors which he felt to be of importance. These were: omission of limbs of family members, size differences, ordering of family members, size differences and expressiveness of facial features, lack of differentiation of any or all figures (especially feeling of self as small and undiffer- entiated), and general expressive features of the lines (as chaotic vs. rhythmic, angular vs. round lines). Although he did not quantify his results, Wolff concluded that the H O . . drawing of the family reveals the child's specific attitude toward each member of the family as well as his wishes and fears concerning them" (p. 1A2). From his total study, Wolff claimed that a unifying concept regarding children's behavior did emerge: "the child's search for self--Who am I? What am I for? The child does not explore the world only in order to gain knowledge, but also to differentiate himself from his environment" (p. xiv). The next study was undertaken by Hulse (1951), who used family drawings as a diagnostic aid in treating child patients. Using the instruction "Will you draw your family (for me)?", he collected "several hundred" drawings from "disturbed" children of all ages. Hulse was looking for signs of oedipal conflicts in the drawings, and felt that the following signs were helpful discriminators: size of figures, relationship to each other, distribution in space, strength of pencil stroke, shading, coloring, omissions, and cartoon effects. He concluded, however, that of even greater significance was the overall gestalt of the picture. Unfortunately, he gives only four examples of this general gestalt factor: the sense of being "up in the air," family members removed to the background, the feeling of being an "outsider," and the feeling of weak— nesses in family members, because of thin arms. Hulse followed up with another study (1952) in which he collected drawings of "normal" children to see if they also revealed family conflicts. An interesting sidelight is that in discussing his project with New York public school teachers, he was informed that a large number of the regular classroom pupils were seen by the teachers as having emotional disturbance, although they had not been referred for help. Accordingly, Hulse collected drawings only from those children the teachers felt were "well adjusted." In his analysis, he compared 120 such drawings with those of his patients mentioned in the study above, and concluded that practically every drawing revealed some form of family conflict. Although he did not quantify the results, Hulse felt that children ". . . project their unconscious conflicts very definitely into their family drawings, and give us a great deal of insight into the family constellation, as it is experienced by the child" (1952, p. 72). In Germany, Flury collected family drawings from children in a psychiatric observation hospital (from 19A3 to 195A), by instructing them to "Draw your family," (Flury, 195A). Again, there is no quantification, but Flury does present 30 cases which illustrate things like theme, size relationships, coloring, posture, placement of figures, omissions, or additions. Flury's purpose for making the study was to find "laws" which could be applied to every drawing variable; for example, that the use of somber colors would invariably be associated with underlying depression, or that chronic masturbators or thieves would always betray themselves by omitted hands in their drawings. He concluded that although these kinds of ". . . peculiarities recur again and again . . .," they were not found to have invariant relationships to any given type of emotional or behavioral problem (p. 12A). In the first reported attempt at objective quantifi— cation, Martin (1955) took off from Wolff and looked at the family drawings of secure vs. insecure children. He tested 75 children, kindergarten through grade two, with the suggestion that each ". . . draw us a picture of his family" (p. 330). The drawings were rated on symmetry, determination of strokes, expansiveness, centering, sharp- ness and distinctiveness of features, activity, and differ— entials between figures on the foregoing items. The data were analyzed by t tests, which revealed only one variable to be of statistical significance in differentiating the secure from the insecure children. It was found that the insecure group was more expansive than the secure group. Unfortunately, however, this was in a direction opposite that predicted by Martin. A second objective study was made by Reznikoff and Reznikoff (1956), with the intent of determining the effect of sex, social class, IQ and race on the family drawing. They asked 100 black and white boys and girls of low and middle income families to draw their family, including themselves. The results indicated that boys omit the mother or her arms significantly more often than girls, and are also more likely to place themselves in the middle of the family group. As compared to middle income children's drawings, low income children made themselves the smallest figure, omitted mother more, drew father without arms, made an older sib the largest figure, and drew the family as if floating in air. The negro children more frequently omitted people's fingers and their own siblings. There was no IQ differential reported by the authors. Their con- clusion was that sex, class and race did have some differ— ential effect on the family drawing, at least in the fac— tors they looked at. The first (and last) objective and quantified study attempting to look at children's adjustment as expressed in their family drawings was done by Lorand (1957). Here, 600 children (kindergarten to grade 9) were asked first to draw a same-sexed person, then instructed "On this page I want you to draw a picture of your whole family. Be sure to put in everyone, including yourself" (p. 21). After 1A minutes, they were asked to label the family members. The drawings were rated for maturity level on the Lorge-Dunn Maturity Rating Scale (In Lorand, unpublished), and the children's adjustment was rated by their teachers on the Haggerty-Olson—Wickman Behavior Rating Schedules (H-O-W), (1930). Lorand found that correlations between rated maturity of family figures and H-O—W scores were all signi- ficant (.01), but small. Girls who drew mother less com— plete were rated as less mature by their teachers, as were boys who drew themselves partially. Also, much lower maturity ratings on the family drawings than on the drawing of a person were negatively related to adjustment. This last finding has an important bearing on the present study because it indicates that (1) children's emotional problems do have some influence on their family drawings; (2) the family drawing tends to be more sensitive to emotional factors than the drawing of a person, at least in children; and (3) the child's emotional problems have a direct rela- tionship to the family scene. The most recent study (Shearn and Russell, 1969) was done at a residential treatment center for disturbed children ages 6 to 12, at the Colorado State Hospital, and introduced the technique of comparing family drawings of the child patient with those of his parents. Several examples were presented which illustrated the dynamic interplay between family members, with similarities and dissimilarities between the drawings analyzed. These researchers found that the instructions to "Draw a family” almost always elicited drawings of the child's own family, yet provided him with some latitude to draw some of his own perceptions of himself in the group which might differ from other's ways of seeing things. Outside of methodological failings and lack of adequate quantification, the major problem with all of these studies seems to be their "hit—or-miss" approach to analyzing the family drawing. There seems to be either the attempt to glean small empirical nuggets, or to strike into the mother lode of some theory of personality which focuses on the dynamics of the individual. Attention given to group or interactional factors has been minimal. Given the family as a highly unified group with its own identity and rules of the game, the writer feels that the potentially most useful way of looking at the family drawing would center more on the dynamics inherent in the unique processes of the self-in—the—group. This approach would consider both individual and group identity, and the reciprocal influ— ences of one on the other. A useful way of looking at this identity reciprocity has been outlined by Erich Fromm (19A1). A System for Analyzing the Family Drawing: Theoretical and Empirical Foundations Fromm: Freedom and fear, the self and others Fromm believes that after a child is born it becomes a biological entity, but that "functionally" it remains one with its mother for some time. These psychological ties are later extended to the family group and the natural world surrounding the child, so that, like primitive man, his deepest sense of personal identity includes ". . the soil he lives on, the sun and moon and stars, the trees and flowers, the animals, and the group of people with whom he is connected by ties of blood" (19A1, p. 50-51). Although these primary ties of extended identity give the individual great security by allaying his fears of isolation and insignificance (the cog-in- the—machine kind of security and significance), they also block his full human growth. They stand in the way of the development of his reason and his critical capacities; they let him recognize himself and others only through the medium of his, or their, parti- cipation in a clan, a social or religious community, and not as human beings; in other words, they block his development as a free, self-determining productive individual" (p. 51). But the growth forces are strong and hard to deny. "Life has an inner dynamism of its own; it tends to grow, to be expressed, to be lived" (p. 206). Fromm speaks of expansiveness and spontaneity of growth as the child grows stronger physically, emotionally, and mentally, and with growing integration of all these parts. He calls this end product of organization and integration the "self" and speaks of the growth of "self—strength" and freedom as individuation (p. AA). However, pro- blems arise in that individuation always brings increased aloneness, and if there have been weakening experiences in the family group, the individual may renounce self and 10 freedom. If this happens, the self is abandoned to a new group, wherein one can submerge one's aloneness in new ties of dependence and functional oneness. The self is thus surrendered by being offered up to an emotional "symbiosis.” New problems are created by these symbioses, and paradoxically enough, because of the very potency of the growth tendency. It seems that if this (growth) tendency is thwarted the energy directed toward life undergoes a process of decomposition and changes into energies directed toward des- truction . . . Destructiveness is the outcome of unlived life. Those individual and social conditions that make for destruc— tion that forms, so to speak, the reservoir from which the particular hostile tendencies-- either against others or against oneself——are nourished" (p. 206-207). Out of the surrendering of self arises three major character types, according to Fromm. These are: (1) Those who feel small and insignificant, who turn their hostile energies against themselves, or elicit punish- ment from superior powers, to which they also tend to cling; (2) Those who feel others are small and insigni- ficant, and cling to them because they need them symbio— tically in order to prove their own greater power by bullying or humiliating them; and (3) Those who withdraw and isolate themselves in a fantasy annihilation of . l everyone else, thus preserving some sense of self. 1These types seem highly similar to Horney's Neuro- tic Trends; moving toward, against or away from people, (Horney, l9A5). 11 These first two character types are extremely dependent on the relationships with those in their groups, and thus would have a high level of group identity. The third type pulls away from people, and thus would have a low level of group identity. As the only alternative to surrender of self, which does not block growth, Fromm sees a new way of relating to the world, but one which comes not from weakness but rather from strength-—strength of self. This new way is a "spiritual relatedness” which is chosen out of freedom, rather than run to out of fear. It consists of ties of caring, involvement, productivity and potency with others-in the world.2 These ideas about human growth might be simplified to the movement of the person from Low individual iden—' tity/High family group identity to High individual identity/Low family group identity. When the individual is young and dependent, because of his feelings of help- less insignificance, his own identity is submerged in that of the family group in such a way that he assumes some of the group's potency and significance. He actually opens up his identity boundaries to encompass the group. 2Perhaps Fromm's spiritual relatedness is best expressed in lines from Robert Frost's The Silken Tent (196A): And its supporting central cedar pole, That is its pinnacle to heavenward And signifies the sureness of the Soul Seems to owe naught to any single cord, But strictly held by none, is loosely bound By countless silken ties of love and fhought To everything on earth the compass round Given parents who are themselves relatively secure as individuals, the child is prodded at the right times, and sheltered when really needed, which enables him to have fuller and fuller experiential intercourse with the world. This results in the growth of a deep and wide experiential fund (of both oneself and of others), with hurts, angers and disappointments, but more importantly with unshakable feelings of worth, significance, potency and caring. Thus, the self becomes "actualized," the group no longer provides the sense of strength needed for growth, and the identity boundaries are closed in about the individual. He is then alone, but with bridges of caring and productive investment spanning the void between himself and cthers. The complications that can arise for the individual in this natural growth process are due to the problems within the family group, which give rise to group sym- biosis or isolation. In the symbiotic group there is an attempt to maintain a very high level of group identity as a protection against fear and change (which includes growth), which causes overdependence in group members. Because of the destructively channeled thwarted growth forces, this overdependence can lead to masochistic sub- missiveness or bullying in the group. In the isolated group, the members feel insecure and untrusting, unloved and separate from others, resulting in counterdependence, 13 low group identity, and destructive feelings directed toward others. These dynamics could be simplified to submerged self/gymbiotic group and isolated self/frag— mented group. Combining high and low group identity with high and low growth patterns yields the following diagram: High Growth Low Growth High family group identity Low family group identity The left column indicates the positive growth pro— cess from early to late childhood, while the right column refers to disruptions of the growth process. Across the top row the child's identity is subordinated to that of the group, while the bottom row reflects the waning influence of the family group identity on the identity of the child. The next step is to fit this theoretical model to the empirical scoring dimensions derived from pilot studies on the family drawing. The Pilot Studies The pilot studies were based on the assumption that a child's family drawing will show significant represen- tations of the child's identity within the larger family group identity. 1A The first pilot was done in the Spring of 1966 by the writer, and consisted of comparing drawings from about 20 third graders in the Haslett Elementary School with about 50 drawings from the files of the Lansing Child Guidance Clinic. Out of this comparison a rudi— mentary scoring system was developed (about ten items), and related to individual and group identity theories. The next pilot study was done in the spring of 1968 with about 75 first, second and third graders in the Portage School system, and about 20 children in Kalamazoo County emotionally disturbed classrooms. The teachers filled out a rating scale developed by the writer which it was hoped gave some indication of the pupil's growth and emotional adjustment. In comparing the adjusted and maladjusted groups, several interesting points came to light. The most significant was that within the adjusted group, about 2 per cent of the first graders placed themselves next to the like-sex parent in the drawing, whereas 90 per cent of the third graders so placed themselves. The second graders were split about 50-50 on this placement. This suggests a very definite change in the child's personal identity system. In addition, not one of the disturbed children (age 9 - ll) placed himself next to the like-sex parent, (that is, exclusive of tucking himself between mother and father). Further, when these children placed themselves within no. A .\ ‘1 “A 15 their sibship, they were more likely to be last, isolated from the others, encroached on physically by others, or submerged in the group. The adjusted kids were more likely to place themselves first, or at the prominent spot in the sibship. In addition to these position fac- tors, other scoring dimensions were worked out to yield the following list: Structure of drawing Position of subject in overall group Position related to each parent Distance-closeness of figures Encroachment-isolation of figures Affectional contact between figures Stance of figures Plane of figures AAAAAAA 1 2 3 A 5 6 7 VVVVVVV Bodies of family members (1) Reality of appearance (as stick figures, cartoon- like, general realism) Size (overall and differentials in body parts) Omissions-additions Facial features, expression Embellishments (Positive or negative; good feeling or demeaning) Sexual differentiation of figures Stereotypy of figures (Degree of variation between figures in attire, stance, posture, expression) ) Action, feeling of aliveness AA AAAA CI) NC“ UTEUUN vv VVVV Also in this pilot study the children were requested to draw a group of their friends as a kind of control on the family drawing. It was found that these drawings differed from the family drawings on many of the above scoring dimensions, thus providing further support for the contention that the family drawing does express some- thing unique about the child in his family group. l6 Integration of Theory and Empirical Scoring: The Scoring Matrix It was at this point that the writer began to see a fit between the results of the pilot studies and Fromm's theory regarding identity, freedom and growth. It was then discovered that the scoring dimensions pre- sented above fit almost perfectly into the previously presented Identity-Growth diagram, High Growth Quadrant l-—Dependence j? A. Position Factors I: "Me" next opp-sex parent g} "Me" lst; sibs by parents p 7' B. Relational Factors Q a Real people a ‘5 Non-demeaned people .3 Positive embellishments E3 Fair to good proportions m a. Close w/o encroachment .c Affectional contact if Some sex differentiation m Few major omissions as follows: Low Growth Quandrant 3—-Overdependence A. Position Factors "Me" in sibship "Me" between parents B. Relational Factors Unreal people (cartoon, etc. Demeaned people Negative embellishments Poor prOportions, all people Overclose, encroachment Non-affectional contact Little or no sex differentiation Major ommissions 17 High Growth Quadrant 2--Independence Low Growth Quadrant A-— Counterdependence m .p E A. Position Factors A. Position Factors .3 "Me" next like—sex parent ”Me" last, sibs by parents +4 ”Me" independent "Me" isolated 04 g B. Relational Factors B. Relational Factors s 0 Real people Unreal people ("Me" only) % Non—demeaned people Demeaned people ("Me" only) 2 Positive embellishments Negative embellishments E ("Me" only) If Good proportions Poor proportions (”Me" only) 3 Optimum distance Isolation ’3 Doing something Turned away, differentness Poor sex differentiation Major omissions ("Me") High sex differentiation No major omissions As can be seen, the scoring dimensions are broken down into Position factors, which are scored by the position of the "Me" in the family group, and Relational factors, which deal more with the relation of the "Me" to the other group members. In looking at the scoring factors, it can be seen that those concerned with Position are much more objective than the Relational ones, and therefore easier to learn, and more economical and reliable to use. The assumption underlying the use of Position is that the "Me" placed next to somebody reflects approach vector(s) to that person, either as they are in reality, or in wishful thinking. These approach vectors can be founded in good feelings (caring, admiration, positive identifica- tion), or in bad feelings (fear, stifled rage, deprivation, l8 guilt, identification with the aggressor, etc.). The Relational factors are concerned primarily with closeness— distance, enhancement—denigration, freedom—boundedness and real-unreal. The application of these two scoring systems to the sample of drawings will be taken up next. Of special interest in terms of developing a screening device will be the comparative validity of the Position vs. the Relational systems. CHAPTER II HYPOTHESES The foregoing theoretical presentation has sug— gested how relationships between individual and family group identity can promote or retard the child's growth of self and emotional well-being. When a child draws a picture of a family, there is evidence that he pictorially and expressively represents the identity and growth dimensions of himself, his family group, and himself-in-the-group. However, there have been no studies with the family drawing technique which have established any norms for use in this way. The purpose of this study, then, can be divided into two major goals: the establishment of some developmental norms for the family drawing technique, using the pro- posed scoring system; and the validation and/or discovery of scoring factors which will differentiate maladjusted children's drawings from those of better adjusted children. The tool to be used is the Scoring Matrix, into which family drawings can be sorted by Position and Relational factors. l9 20 In the Scoring Matrix we have the two independent variables of age (or grade), and adjustment. The recip— rocal of Family group identity/Individual identity is assumed to be negatively associated with age, and growth to be positively related to adjustment. The four quad- rant Matrix can be assessed across the Identity or Growth dimensions alone, or in terms of the interactions between these dimensions via the individual Quadrants. The hypotheses to be used are designed primarily to test the validity of the sorting techniques in terms of a straightforward "fit" with the real world. If the sorts are valid within the framework of the Scoring Mae trix,then we would predict the following results: Younger, normal children would tend toward Dependence, with older, normal children shifting to Independence. Younger, emotionally disturbed children would tend toward Over- dependence, whereas the older disturbed kids would move more toward Counterdependence. Disturbed children should be proportionately higher than normals on Overdependence and Counterdependence, and increasing age should not change this picture. The hypotheses, then, should be predictive of these results. Hypothesis I: The family drawing of a child will vary in its placement on the Scoring Matrix as a function of the child's age and adjustment. Hypothesis Hypothesis Hypothesis Hypothesis Hypothesis Hypothesis IA: IB: II: IIA: IIB: IIC: 21 The drawings of younger, adjusted children will tend to cluster most in the High group identity-High growth Quadrant of the Matrix. With increasing age of subjects, the family drawings will tend to move from the High group identity- High growth Quadrant of the Scor- ing Matrix to the Low group identityeHigh growth Quadrant. The family drawings of malad- juSted children will differ from that of adjusted children, (con- trolled for age, sex, social class and race). The maladjusted children's drawings will tend to cluster on the Low growth side of the Scoring Matrix. With increasing age, the malad- justed children's drawings will move from the High grogp identity row to that of Low group identity, but not as much as the movement of the adjusted children's drawings. With increasing age, the malad- justed children's drawings will tend to remain within the L93 growth column of the Matrix. CHAPTER III Children in two elementary schools and eight ungraded rooms for the emotionally disturbed made draw— ings of "a family," which were sorted by (1) Position determinants and by (2) Relational determinants. Within each of these sortings, the drawings were segregated into normal and disturbed samples, which were analyzed by age (or grade), sex, social class, and race. Subjects There were 5A0 drawings in the total sample. It was necessary to discard 119 because of lack of self— representation, omission of parents, or other defects. The normal group lost 100 drawings and the disturbed group lost 19, leaving A56 normal and 8A disturbed drawings. Normal Group Of the A56 normal drawings, 222 came from Edison Elementary School, and 23A from Milwood Elementary, both in the Kalamazoo City School System. There were 215 boys', and 2A1 girls' drawings; grades one through six, 22 23 ages six through thirteen. (See Appendix C for breakdown by school, age, grade and sex). Emotionally Disturbed Group The overall emotionally disturbed sample of 8A, made up of 66 boys and 18 girls, came from two major sources: Al from special education classrooms, and A3 from the regular classrooms, via teacher rating. The Kalamazoo emotionally disturbed classrooms are made up of pupils who present severe problems of manage— ment or withdrawal in the regular classroom, but who do not meet the requirements for the special classes for "perceptual disturbance"-—(an educational euphemism for brain damage). There are only four classrooms for the disturbed child in Kalamazoo County, and the classes are kept under 10 pupils each, so they probably contain some of the more severely disturbed children in the school system. The writer visited these classrooms, and felt that a high proportion of these children would be diagnosed as schizophrenic. Because there were only 26 drawings from the Kala- mazoo classrooms, the disturbed rooms in Grand Rapids were also tapped, for another 1A drawings. Teachers in the regular classrooms identified the children they thought emotionally disturbed, and were reminded not to overlook the extremely overdependent child. (See Appendix C for a breakdown of the emotion- ally disturbed sample by source, sex and age.) 2A The disturbed drawings from the two sources were combined into a single sample because their scores were similar in the Scoring Matrix. N Sex, Age, Grade, Race, Social CI'S‘ r C The sex of the subject was deduced from his first name, while age and grade were self—reported. Black subjects' drawings were identified by their teachers. The Social Class variable was determined primarily by the location of the schools involved in the study. The middle class Milwood School is located in census tracts with an average home value in 1960 of $1A,560, whereas the average home value for the lower class Edison School tracts was $9,330 (Bennett, 1961). The range of home values in Kalamazoo at the time (the tract medians) was $7,700 to $19,300, ($12,000 median) so that the middle class and lower class designations seem appropriate for the two school districts. One of the Kalamazoo emotion- ally disturbed classrooms is located in the Milwood School, and another in the Lexington Greens School (an area of the same level homes as Milwood). The other two disturbed classrooms are in the Comstock School System, where the average home value is $8,300. The teachers of the dis- turbed classrooms in Grand Rapids identified their pupils as middle or lower class. 25 Collecting the Drawings Administration The participating teachers administered the drawing test to their classes on the basis of written instruc— tions (see Appendix B), and using 8—1/2 x ll—l/2 white paper with a place provided for pgme, Egg and gaggg in the upper right. Among the few reports of administrative problems were: one White boy who drew a Black family, then balked "me," but compromised by at labeling one of the members writing "not-me" over a like-sex child; a suspicious group in a disturbed classroom in Grand Rapids who felt they were being tested, so made the drawings on the back of the paper, and omitted the "me"-—(a11 of this by a note from the teacher, who seemed a little suspicious herself). The writer was also the object of an hour-long telephone harrangue from a mother who had gone to the principal and demanded her son's drawing, because she suspected that the school superintendent might use it against her. She seemed to fear that something was wrong with her son (which his teacher confirmed), but which she could not own up to. 26 Scoring The identifying data were masked on the drawings, which were shuffled, then sorted by Position and Belg— tional determinants. After the two sortings, the draw- ings were divided into normal and emotionally disturbed groups, which were analyzed separately. Position In the position sort, the drawings were first assigned to one of eight possible positions of the "me" in the family group, which determined the second step of placing the drawings in a Quadrant of the Scoring Matrix on the basis of its position, as follows: High Growth Low Growth High Group Next opp—sex parent In sibship Identity lst; sibs by parents Between parents Low Group Next like—sex parent Last; sibs by parents Identity Independent Isolated Oppositional Sorting the drawings for all but three of these determinants is objective, and therefore quite easy to H H learn. The three subjective determinants are the me independent, which requires that the "me" have distance, but is doing something (there may be intervening objects, but they are part of the scene); "me" isolated, judged by H H distance, but the me not doing anything, or with inter- vening objects which are instrusions in the scene, or 27 H H with walls in the way; and me oppositional, which shows conflict, or hostile type behavior. Two independent judges agreed .8A, .72 and .80 respectively on these three subjective placements, with the overall agreement being 79 per cent. Relational Sort Two judges independently sorted the drawings on the basis of the Relational determinants, by assigning each drawing to one of the quadrants of the Scoring Matrix. About a third of the drawings were found to have high clustering of these determinants, and thus fit easily into one of the four quadrants. Another third had suffi- cient clustering of factors to provide strong clues to Quadrant placement. The remaining third showed such slight clustering of the Relational determinants as to necessitate "forcing" into a Matrix Quadrant. Agreement between the judges' sorts averaged 73 per cent. Sorting was done in a series of six sub—groups, with consecutive agreements of .69, .73, .73, .73 and .80. The judges together made the final placement of the 27 per cent of drawings disagreed on in the indepen- dent sort. In attempting to sort the drawings according to the Relational factors in the Scoring Matrix, it soon became apparent that the judges were assigning drawings to Quadrants by the following criteria: 28 Real people Non-demeaned people Closeness w/o encroachment Possible sterotypy, but w/o negative features Real people Non-demeaned people Distance w/o isolation No stereotypy Positive embellishments No major omissions "Group Gimp" "Group Mash" Demeaned people Stereotypy, with negative features Isolation of "me" by: Distance Barriers Different planes Disproportionate sizes "me Gimp" "Group Gimp" refers to the sharing of similar defective or demeaning qualities by all group members, while the "mash" refers to the jamming together of group members to give a mash-like quality to the drawing. The "me gimp" refers to the negative, demeaning or defective qualities applied only to the "me"; but there may be a reversal in terms of everyone but the "me" being "gimped.“ CHAPTER IV RESULTS The Position and Relational sorting techniques are ways of placing the family drawings into quadrants of the Scoring Matrix. The Matrix has two major dimensions, Family Group Identity and Growth, both of which are divided into higp and lpfl_quadrants. The resulting four- quadrant Matrix can be assessed across the Identity or Growth dimensions alone, or in terms of interactions between these dimensions. The major variable theorized to affect Family Group Identity alone is that of age; i.e., "younger" children being associated more with high family group identity, and "older" children more with low family group identity. The hypotheses dealing with age changes and placement in high and low group identity rows, then, test the validity of the sorting techniques along this dimension. The major variable associated with Growth alone is that of adjustment, with the prediction that "normal" children will be associated more with high growth, "dis— turbed" kids more with low growth. The hypothesis con— cerned with placement of normal and disturbed drawings 29 30 in high and low growth columns respectively, is intended to test the validity of the sorting techniques along this dimension. The hypotheses dealing with placements in one of the four quadrants of the Matrix concern interactions of identity with growth, and are evaluated by analyzing age and adjustment. These are more subtle tests of the valid— ity of the sorting techniques in terms of their power to discriminate interactive effects of the major dimensions. The hypotheses listed under I concern the normal sample, whereas those under II compare the disturbed and the normal samples. Position Analysis Hypothesis I: The family drawing pf g child will vary 13 its placement pp the Scoring Matrix pg g function pf the child's age and adjustment. This is a general hypothesis which implies that the scor- ing system should be sensitive to age differences in the drawings of normal children. For the purposes of this analysis grade rather than age was used, because of the greater experiential similarity of children within grades, and the cleaner computational possibilities offered by six discrete grade levels as opposed to seven non-dis- crete age groupings. The results for this hypothesis appear in Figure l, for boys and for girls. It can be seen that there are 31 some differences in quadrants by ages, although there is a reversal in quadrants l and 2 between boys and girls. Chi squares for these data are given in Table 28, quad— rants 3 and A being combined because of small frequencies at later grades (see Appendix F, Table 28). The chi square for boys is significant at the .01 level; that for girls is nonsignificant. Boys ,’/ Q2 P 70~ /,/ E 60- ,’ / R 50'- I/ no- ,/ '91 230— ,»-.-—__-s-—--——’ 201 W~ N ,7 \-.-‘ _:*WQ3 T 1.0-1 I '\‘_______.-—'Qu O T l l I l I I II III IV V VI 70 _ Girls P 60 , §5o_ ’,L___~- A0 _ .,- ’ c 30 - . E 20 ~ m- - N 10_, r, s” T o \\\\\‘\ I l ' I l | 7 I II III IV V VI Figure 1. Per cent of boys and girls in quadrants at each grade: Position analysis. 32 From inspecting Figure 1, it can be seen that the younger boys tend to quadrant l, but with increasing age, they tend more toward quadrant 2. Quadrants 3 and A start out with medium percentages of younger boys, and both decline in frequency with increasing age. The girls are consistently high in quadrant 2 at all grades except grade V. The reversal of quadrants l and 2 for boys vs. girls suggests that the sorting tech- nique differs for the two sexes. Hypothesis IA: The drawings pf younger, adjusted children will tend to cluster most ip the High Grafip Identity/ High Growth quadrant pf the Matrix. This is a more specific hypothesis which says, in effect, that when normal children are younger (Grades I- III), their family drawings will show more dependent qualities as opposed to qualities of overdependence, counterdependence or independence, and accordingly will tend to be sorted more frequently than chance expectation into quadrant l of the Scoring Matrix. Figure 1 indicates that for boys this seems to be the case, and the chi square confirms this, based on a 25 per cent probability that any given variable will be found in any single quadrant, (See Table 2). However, the hypothesis is not confirmed for girls, which is obvious both in Figure l and Table 2. 33 TABLE 2.--Observed and Expected Frequencies of Family Drawings in Position Analysis: Grades I~III, Quadrant 1 vs. Quadrants 2, 3 and A. Sex Quadrant 0 E x2 p BOYS 1 51 30.75 (25%) 6 85 01 2,3,A 72 92-25 (75%) ' ' Girls 1 27 31 (25%) 2,3.A 97 93 (75%) '56 NS As was suggested from the results of Hypothesis I, there appears to be something in the Position sort which does not hold up for girl's drawings placed in quadrants l and 2. Hypothesis IB: With increasinga _g_ ofs subjects, the family drawings will tend to move from the High Group Identity/ High Growth quadrant of the Scor— __g Matrix to the Low Group Identity/High Growth quadrant. This hypothesis indicates that as children get older (grades IV-VI), their drawings will show more independent qualities when compared to the younger drawings' qualities of dependence, and that there will be a shift from quad- rant 1 to quadrant 2, accordingly. Figure 1 shows this trend for boys, but not for girls, who were high in quadrant 2 throughout all grade levels. Table 3 indicates the hypothesis supported for the boys, but not for the girls. 3A TABLE 3.—-Position Analysis of Family Drawings: Grades l—III vs. Grades IV—VI, Quadrant 1 vs. Quadrant 2. Sex Quadrant Grade I-III Grade IV—VI x2 p Boys % 2% 33 5.18 .05 Girls 1 27 36 o 2 56 56 '56 N” All of the hypotheses under I were confirmed for boys, suggesting validation of the Position sort for them. However, none were confirmed for the girls, indicating that the sorting criteria should be reassessed for them. Hypothesis IIA: The maladjusted children's drawings will tend to cluster pp the Low Growth sIde pf the Scoring Matrix. This is to say, the emotionally disturbed drawings will show up more in quadrants 3 and A than in quadrants l and 2 when compared to the distribution of the normal drawings. This hypothesis concerns the Growth variable with Group Identity held constant, making the prediction that low growth will be associated with emotional dis- turbance. Figure 2 shows the proportion of drawings of nor- mal and disturbed children sorted into the High Grthh half of the Scoring Matrix; (the curves for the Low Growth half would simply be the reverse of these). Age 35 is used here rather than grade because the children in the special education classrooms for the disturbed are ungraded. L___~ Normal r--«~ Disturbed 9O _ f 80 ‘1 ’ \\ / P I \ / E 70 ’ \ ,\ R 60 I M // \\ .7 / \k‘ / \\ ..‘/ \ C 50.3 // \ E A . N 0 s T 30,1 20 '7 10., 0 l l r l I 7 8 9 10 ll 12 & 13 AGE Figure 2. Per cent normal and disturbed in quadrants 1 & 2: Position analysis. 36 The overall hypothesis is supported at the .05 level (Table A). Table A also contains the chi squares for the variables of sex by social class, and race, in order to assess tie effects of these factors on the sorting results. TABLE A.-—Position Analysis of Family Drawings by Sex, Class and Growth. Factor Growth Normal Disturbed x2 p Overall High 329 51 MC Boys High 8A 13 Low 2A 11 6°90 01 MC Girls High 103 A Fisher's Low 27 A Test .006 LC Boys High 70 30 , Low 35 12 '13 NB LC Girls High 72 A Fisher's Low Al 6 Test NS Blacks High 25 5 Fisher's Low 13 3 Test NS It appears that a disproportionately low number of dis— turbed lower class boys' drawings were sorted into the Low Growth half of the Matrix. There were also a high number of normal lower class boys' drawings placed in the Low Growth column. This last effect also shows up for normal lower class girls, which suggests a social class factor in the sorting technique. That is, when compared 37 to normal, middle class drawings, the normal, lower class children's drawings tend to be more like those of the disturbed children. In addition, drawings of disturbed lower class boys tend to be like those of all normal boys, with the effect probably in quadrant 1, older boys H who draw the me" next to the mother. The black children, who are all lower class, follow much the same pattern. In Table 5 the results of the age levels for Hypo- thesis IIA (ages pooled because of low frequencies in the disturbed group) suggest that the drawings of younger normals tend to be disproportionately placed on the Low Growth side of the Matrix. TABLE 5.—-Position Analysis of Normal and Disturbed Family Drawings: Age and Growth. ‘2 Age Growth Normal Disturbed x p 6 & 7 High 76 A Fisher's Low A0 5 Test NS 8 & 9 22%“ 2% 1g 2.A2 s 10 a 11 Eggh 1%; 13 2.21 NV 12 a 13 Eiéh Ag i3 9,53 01 Following up the apparent discrepancy in the normals by age, Table 6 shows that younger normals do tend 38 to be disproportionately placed in the Low Growth column, as compared to older normals. TABLE 6.——Position Analysis of Normal Family Drawings: Sex, Age and Growth. Sex Growth Age 6-9 Age 10-13 x2 p Boys High 81 73 Low A5 1“ 8.93 .01 Girls High 9A 81 Low A8 20 5°06 '05 Hypothesis IIB: With increasing egg, the maladjusted children's drawings will move from the High Group Identipy row to that of Low Group Identity, but not ee “much as the movement of the adjusted children's drawings. Here, Growth is held constant and Group Identity is varied, with the prediction that the High Group Identity row (quadrants l and 3) will contain more younger child- ren's drawings, and the Low Group Identity row (quadrants 2 and A) more older children's drawings; the effect being more pronounced for normals than for emotionally dis- turbed. Although Figure 3 indicates a slight trend in the hypothesized direction for the normals and disturbed, neither reaches statistical significance (see Appendix F, Table 29). 39 '————4 Normal “"”“ Disturbed 90 _ 80 a 70 a 60 d 50 3 A0 30 d 20 _ 10 621110 :Ul'liltd Figure 3. Per cent in High Group Identity Rows, normal and disturbed: Position analysis. In breaking the results down by age, the normals show the hypothesized trend at age 10 and 11, and the disturbed at age 12 and 13, but these do not reach significance (Table 7). AO TABLE 7.--Position Analysis of Normal and Disturbed Family Drawings: Age and Group Identity. Age Adjustment Group 6&7 8&9 10&11 12&13 2 p Identity X Normal High 52 77 56 26 Low 5A 79 83 23 3'72 NS Disturbed High 6 13 16 11 Low 3 12 9 1A 2'57 NS Hypothesis IIC: With increasing age, the malad- Justed children's drawings will tend 32 remain within the Low Growth column of the Matrix. This hypothesis is based on the assumption that increasing age of the disturbed children should not influ— ence their drawings towards movement out of the Low Growth column. Figure 2 and Table 8 offer support for the hypothesis. It should be noted that significant chi squares here indicate an actual increase for the older disturbed frequencies in the Low Growth column. Those variables that did not reach significance do not, on inspection, show a decrease in frequency in the Low Growth column with age increase, and thus do not represent a reversal of the hypothesis. Al TABLE 8.--Position Analysis of Family Drawings for Older Children (Age lO-lA): Adjustment and Growth. Growth Factor Adjustment High Low x2 p Overall Normal 15A 3A Disturbed 29 21 11°35 '001 MC Boys Normal 37 5 Disturbed 5 10 1A.38 '001 MC Girls Normal A0 8 Fisher's Disturbed l 1 Test NS LC Boys Normal 36 9 1 Disturbed 22 5 ‘02 AS LC Girls Normal Al 12 Fisher's Disturbed l 5 Test .005 Blacks Normal 11 5 Fisher's Disturbed 3 1 Test NS Relational Analysis The hypotheses for this analysis are identical to those in the Position analysis. As there, the I hypo— theses deal with the normal sample, whereas those under II are concerned with the normal vs. disturbed samples. Hypothesis I. The plot of the four quadrants by sex and grade (Figure A) shows that there are differences by age and quadrant. Small frequencies in quadrants 2 and A led to grouping quadrants for the chi square (Table 9), where both boys and girls attained significance. 100 90 80 7O 60 50 A0 30 2O 10 *U r. -1 ZUE. I—EZEO 100 90 80 7O 60 50 A0 30 20 10 WWW HZETJO Figure A. A2 Boys 1 T ---—— I I I I III IV V VI GRADE Girls .1 fl /k~'~r’"-—' Q2 “ - Q3 0—- ______ \___....—o——-—-' Q“ I I I III IV' ‘V ‘VI GRADE Per cent of boys and girls in quadrants by Relational analysis. A3 TABLE 9.--Relational Analysis of Family Drawings, Grades I-VI: High vs. Low Growth. Grade Sex Growth I II III IV V VI X2 p Boys High l9 19 26 25 15 21 / Low 31 16 12 7 8 1A 10°05 '02 Girls High 23 28 3O 35 31 38 Low 23 ll 9 3 6 6 26.67 .01 Hypothesis IA As mentioned above, quadrant 3 contains dispropor- tionately high numbers of the younger children, however, Table 10 reveals that the younger children (Grades I—III) cluster significantly in quadrant l, with the effect being greatest for girls. TABLE 10.-—Relationa1 Analysis of Family Drawings in Grades I-III: Quadrant 1 vs. Quadrants 2, 3 and A. Sex Quadrant 0 E x2 p Boys 1 63 30-75 (25%) 9 2,3,u 60 92.25 (75%) 16'83' '001 Girls I 81 31 (25%) 39.13 .001 2,3,“ A3 93 (75%) AA Hypothesis IB Figure A shows the trend for boys and girls in Grades IV-VI to decrease in frequency in quadrant l and to increase in quadrant 2. The chi squares in Table 11 are significant for both sexes. TABLE ll.--Relational Analysis of Family Drawings: Younger vs. Older, Quadrant 1 vs. Quadrant 2. Sex Quadrant Younger Older x p Boys 1 63 52 2 l 9 5.70 .02 Girls 1 81 8A 2 0 20 15.53 .001 Hypothesis IIA Figure 5 shows the percent of disturbed and normal drawings sorted into the High Growth column of the Matrix, with the normals predominating on the High Growth side, (and the converse being the predominance of disturbed on the Low Growth side). Although chi squares for the sex by class variables all reach significance (see Appendix G, Table 30), that for the black children does not. The problem is not with the sorting of the disturbed drawings as much as with the normals; that is, roughly the same number of normal drawings were sorted to the right column as to the left. “5 This suggests that the Relational sorting technique either does not discriminate black normals, or it turns up something in these drawings that is akin to the drawings of disturbed children. An alternative is that more black "normals" have incipient emotional disturbances than their white peers. r—————4 r-—-—-4 100 0 3.: 70.. 60 _ 50 A0 a 30- IO SCU'U *3sz Normal Disturbed Figure 5. Growth column of Matrix: Age and per cent normal and distrubed, High Relational analysis. A6 It should be noted that lower class boys show a reversal, in that normals predominate over disturbed in the Low Growth half of the Matrix. This is identical to the effect found in the Position analysis. Table 12 shows support of the hypothesis at all age levels except the youngest. As in the Position analysis the younger normal children tend toward placement on the 2 Low Growth half of the Matrix, (x of 5.69 for boys, 12.69 for girls, see Appendix G, Table 31). TABLE l2.-—Relationa1 Analysis of Family Drawings and Age: High vs. Low Growth. Age Growth Normal Disturbed x2 p 6 & 7 High 58 2 Fisher's Low 58 7 Test NS 8 & 9 Sigh 133 23 25.52 .001 10 & ll Eiih 1&2 2% 6u.16 .001 12' 13 E25?" $8 3 Hypothesis IIB The hypothesized trends of movement from the High Group Identity to the Low Group Identity rows with increas- ing age are seen in Figure 6 for both normal and disturbed children's drawings. However, the only chi squares that A7 reach significance (see Appendix G, Table 32) is that for the normal middle class girls; (Fisher's Exact Test yielded a probability of .06 for the normal middle class boys). Figure 6. WWW! HZFJO '“' Normal o--—o Disturbed 100 90 - 80 70 - 60 50 L A0 30 2O 10 1 7 8 9 10 11 12 13 AGE Per cent in High Group Identity row, normal and disturbed: Relational analysis. A8 Table 13 shows that the age breakdown for normal boys is not significant, although normal girls reach the .05 level, with the major effect at age 12 and 13. (The disturbed sample could not be assessed by age level because of small frequencies). It should be noted from Table 13, that although the boys show much the same pattern as the girls at the 12 and 13 year level, the results at other ages tend to obscure this effect. TABLE l3.-—Relationa1 Analysis of Normal Family Drawings and Age: High vs. Low Family Group Identity. Age Group 2 Sex Identity 6&7 8&9 10&ll 12&13 X p Boys High A0 58 A9 18 Low 11 12 13 7 1'58 NS Girls High 58 68 65 16 Low 7 9 13 8 8.06 .05 Hypothesis IIC Figure 2 and Table 1A support the hypothesis that the older disturbed children's drawings will tend to remain in the Low Growth column of the Scoring Matrix. Although the black variable does not reach significance, it does not show a reversal of the hypothesis. A9 TABLE lA.-—Relational Analysis of Family Drawings, ages lO-lA: Normal vs. Disturbed, High vs. Low Growth. Growth Factor Adjustment High Low x2 p Overall Normal 1A9 39 Disturbed 8 A2 67°32 '001 MC Boys Normal 3A 8 Disturbed 2 l2 17'5“ '001 MC Girls Normal A0 5 Fisher's Disturbed 0 2 Test .02 LC Boys Normal 26 19 Disturbed 5 23 9°68 '01 LC Girls Normal A9 7 Fisher's Disturbed l 5 Test .001 Blacks Normal 12 5 Fisher's Disturbed 2 2 Test N3 Age trends applicable to this hypothesis may be seen in Table 5. Here it is seen that from age 8 on, the dis— turbed show up more in the Low Growth column, and peak at age 10 and 11, proportionate to normal children's drawings. Additional Findings Following are some general and specific findings derived from the results of the two sorts. For the Posi- tion sort, there is a general look at the distribution of the variables in the Scoring Matrix, and a closer look 50 at the eight possible positions of the "me" in the family group. The latter will have some theoretical implica- tions for the factors of identity and growth. For the Relational sort, the Matrix distribution of the variables will be considered. Finally, the Position and Relational sorts will be compared in terms of percent agreement of quadrant placement for individual drawings, and any simi— larities in outcome of the sorts. Position Sort In looking at the High vs. Low Growth Columns, it can be seen that the middle class are more on the High Growth side of the Matrix (x2 = 9.56, Appendix H, Table 33). The blacks, who are all lower class, follow the same pattern. The drawings of normal children are not equally distributed over the four quadrants. Most fall into quadrants l and 2, with relatively few in 3 and A (Appendix H, Table 3A). Boys and girls are also distributed differentially, with more boys in quadrant l, and more girls in quadrant 2 (see Table 15). Grades I-IV account for these dif- ferences (see Appendix H, Table 35). 51 TABLE 15.—-Family Drawings in Each Quadrant: Position Analysis of Normal Boys and Girls. Quadrant Boy Girl X2 pa 2,3,1 2; 133 10.78 .02 1,3,i II; 3:? 9.59 .02 1,2,3 33 $1? .72 NS 1,2,3 32 $83 ‘ .96 NS aTwo-tail test In comparing the distribution of normal vs. dis— turbed children's drawings, quadrant 2 approaches significanCe with a greater proportion of normals, as does quadrant A with more disturbed (see Appendix H, Table 36). Quadrant 3 is found to contain proportion- ately more disturbed girls as compared to boys, (x2 = A.A8, see Appendix H, Table 36). Each quadrant in the Position sort is made up of two of the eight possible placements of the "me" in the family scene. The most predominant position was the "me" drawn next to one of the parents, accounting for 58 per cent of the boys' drawings, and 56 per cent of the girls'. Of the boys, 38 per cent were next to the mother, 20 per cent next to the father. The girls had 35 per cent 52 next to the mother, 21 per cent next to the father. Except for the last, these frequencies are above chance expectation, (respective chi squares of 3A.32, 8.03 and 32.68, Appendix H, Table 37). Boys outnumber girls in placement next to the opposite-sex parent 2 = 15.26), (X shereas l6. like—sex parent (x2 (see Appendix H, Table boys and girls tend to mother rather than the 65). draw the father. girls outnumber boys next to These differences 38) simply indicate that both "me" next to the Figure 7 shows that boys start out in the lower grades next to the mother, but grade V brings a reversal, with the placement next to the father coming to the fore. The girls start out next to the mother, reverse at grade V, but return to the same pattern in Grade VI. 90 A Boys Girls 80 3 70 _ 60 - 50 - By Mother A0 d / By Mother 30 _ \//\‘X /\\‘\/\ 20 - ,,\ By Father ,I By Father / ./x<:: lO _ " ‘1‘”‘““*“"J ”’I“\*\ .,r/ \r" 0 I I I 1 I I r 7r 1 r T r I II III IV V VI I II III IV V VI Figure 7. Per cent of normal boys and girls drawn next to each parent in grades I-VI. 53 An interesting social class effect is shown in Table 16. The lower class show a nearly significant trend in fewer placements next to mother in grade I. When placement next to either parent is analyzed, the effect becomes significant. This suggests less close— ness between the lower class children and their parents at younger ages. TABLE l6.-—Family Drawings in Grade I with "me" Next to Parents: Middle vs. Lower Class. 2 "Me" placement MC LC X pa Next to mother 25 11 Others 26 3” 5.16 NS Next to parents 36 17 Others 15 28 9°12 '02 aTwo-tail test In comparing normal with disturbed on the "me" placement, normals are significantly higher on Indepen- dence (x2 = 9.11), the disturbed on Isolated and Oppo- sitional (x2 of lA.65 and Fisher p = .0005, see Appendix H, Table 39). When the significant placements are broken down by sex, the same relationships hold, as given in Table 17. 5A TABLE l7.—-Family Drawings and Placement of "me": Normal vs. Disturbed Boys and Girls. Sex "Me" placement Normal Disturbed x2 pa Boys Independent 2A 0 Others 189 66 6'76 '01 Girls Independent 27 0 Fisher's Others 216 18 Test .001 Boys Isolated 7 7 Fisher's Others 206 59 Test .001 Girls Isolated 15 3 Fisher's Others 239 15 Test .001 aTwo-tail test Relational Analysis In the Growth columns in the Relational sort results, both girls and the middle class tend toward High Growth, whereas boys and the lower class tend toward Low Growth, (x2 of 15.08 and 28.91, see Appendix I, Table A0). The social class result is in agreement with that found in the Position analysis. Regarding the distribution of the variables by quadrant, it was found that: (1) Girls dominate in quad- rant 1, boys in quadrant A; (2) middle class predominates in quadrant 1, lower class in quadrants 3 and A; and (3) blacks are dominant in quadrant A, (see Appendix I, Tables A1, A2 and A3). 55 A noteworthy trend for the black children is for the normals to be low in quadrant 3 (Overdependence), but significantly high in the Counterdependent quadrant A, (x2 of 6.A2, see Appendix I, Table AA). When comparing the normal and disturbed groups across the sex, class and race variables in each quadrant, an interesting outcome is that in quadrant A there are no significant differences between normal and disturbed for any of the variables, whereas there are differences in the other two quadrants, (Appendix I, Table AA). There is the suggestion that "normal" drawings assigned to this quadrant are perhaps more like their disturbed counter— parts than in any other quadrant. For children, isolation seems to be the most drastic technique for interpersonal adaptation. Cogparing the Position and Relational Sorts The analysis of the two sorting techniques will be wound up with a comparison between them. Agreements between Position and Relational results would offer fur- ther support of the validity of the techniques as well as their theoretical underpinning. The per cent agreement between the two sorts by quadrant is as follows: u2% l 27% 56 The overall agreement is 33 per cent. Chance agree- ment would be 25 per cent. It should be noted, however, that with a few adjustments to the Position analysis (see Discussion), the agreement is raised to 53 per cent. Other agreements between the Position and Relational analyses which relate to the variables of sex, social class and race are: the tendency for normal, middle class toward placement in the High Growth column, and lower class in the Low Growth column of the Scoring Matrix-—particularly lower class boys; the tendency for girls to predominate in quadrant 2; and the "black factor" of black loading in quadrant A (Counderdepen- dence). An important similarity is the tendency for the youngest normal children's drawings to show up in the Low Growth column of the Matrix——particularly in quad— rant 3, Overdependence. Apparently both sorting tech- niques are sensitive to something in young children's drawings which makes them similar to the disturbed drawings, but which cannot be related to the Growth dimension as it is defined in the Scoring Matrix. But more on this in the Discussion. CHAPTER V DISCUSSION Position Analysis—~Th Hypotheses Sorting by Position assumes that the placement of the "me" in the drawing of a family can be a highly mean- ingful communication about a child's identity within his family group identity. There is the second assumption that placing the "me" next to someone suggests an approach vector directed at that person, which exists behaviorally or in fantasy. The approach vector can be positive, in terms of growth needs like caring, admira— tion, and positive identification; or negative as in overdependence, fears, identification with the aggressor and anger; or ambivalent with positive and negative mixtures of feelings. Finally, there may be vectors of avoidance toward others expressed by isolation of the "me." With these things in mind, results in the Posi— tion analysis which failed to support the hypotheses are explainable in terms of two gross errors in the assump— tions underlying the scoring system. The first is that the Opposite and Like-sex parents mean the same thing to both boys and girls, (i.e.; that boy next to mother is 57 58 analagous to girl next to father, and vice versa). The other is that the "me" placement means the same thing at every age or grade level. The effect of these assump- tions will be brought out in considering the non—support given some of the hypotheses. The Hypotheses (IA and IB) associating younger normals with quadrant 1 (dependence) and older normals with quadrant 2 (independence) were supported for boys but not for girls. The association of normals with the High Growth and disturbed with the Low Growth columns of the Scoring matrix (Hypothesis IIA) was supported for all variables except the youngest, lower class and black children. The hypothesized placement of younger children in the High Group Identity row and older kids in the Low Group Identity row of the Matrix was not found generally. Finally, the IIC Hypothesis that placement of disturbed drawings in the Low Growth column will not decrease with increasing age was supported for all groups other than the disturbed black children. Regarding the apparent lack of association of younger and older normal girls' drawings with dgpendence and independence respectively, the problem appears to be in the meaning of the "mg" £353 33 the like-sex parent. Younger girls drawing "me" next to mother probably fits with dependence rather than independence as was postu- lated in the scoring system. Acbordingly, if the 59 younger normal girl's drawings (grades I-III) are moved from quadrant 2 (independence) to quadrant l (dependence), then the I, IA and IB Hypotheses are all supported for girls as well as boys. The Hypothesis (IIA) predicting clustering of disturbed drawings in the Low Growth column of the Matrix was not supported for the lower class, black or youngest groups. The major problem for lower class boys appears H H to be with older boys who place the me next to mother. These drawings were sorted as dependent, whereas older boys next to mother undoubtedly should be considered overdependent. If the age lO-lA boy's drawings are moved from quadrant l (dependence) to quadrant 3 (overdepen— dgngg), then the Hypothesis receives significant support for the lower class boys--primarily because this change would move eighteen disturbed drawings from the High to the Low Growth column of the Matrix, (see change #2, dia- gram p. 61). It had also been assumed that the "mg" last, sibs next t2 parents position indicates counter— dependence. Although this may be true for boys, it does not hold up for girls. Perhaps this position for girls is one of a counterbalance to the parents in the family, as in the "little mother" role, particularly for the lower class. This placement of the "me" is found in none of the disturbed girl's drawings, whereas there are four disturbed boys (age 10 and over) so placed. 60 If the drawings of children ages 6-9 are moved from the counterdependent quadrant A to the dependent quadrant l, Hypothesis IIA becomes significant for lower class girls, also, (see change #3, diagram, p. 61). The fourth questionable Matrix assignment in the Position scoring system is that of the " me" in sibship for younger children. The bulk of the normal children's drawings in the Low Growth quadrant 3 (overdependence) are accounted for by this placement. If these drawings for ages 6—9 are moved from quadrant 3 to quadrant l (dependence), then Hypothesis IIA becomes significant for normal black children, although there is still the overloading of the black normals in the counterdependent quadrant A, and also for the youngest age group, (see change #A, diagram, p. 61). The Hypothesis (IIB) associating younger and older children respectively with High and Low Family Group Identity would receive significant support as a result of the changes outlined above. The last Hypothesis (IIC) which indicates that increasing age of the disturbed should not reverse their tendency to cluster on the Low Growth side of the Matrix was not supported for black children. This is largely due to the dominance of quadrant A (counterdependence) by black children, and especially younger ones. This suggests a factor of counterdependence and/or isolation 61 for younger normal black children which will be elaborated 1n the Discussion of the Relational Sort analysis. The original Position Sort system and the modifica- tions outlined above as they pertain to Hypotheses IIA, IIB and IIC are given in the following diagram. The altered placements are numbered to clarify the changes. Position Sort Scoring Dependence Overdependence "Me" next opp—sex parent* (2) "Me" between parents "Me" lst, sibs by parents "Me" in sibship* (A) Independence Counterdependence "Me" next like-sex parent* (1) "Me" last, sibs by parents* (3) "Me" independent "Me" isolated "Me" oppositional Modified Position Sort Scoring Dependence Overdependence "Me" next mother* (1 & 2) "Me" between parents (ages 6—9) "Me" last, sibs/parents* "Me" in sibship* (A) (3) (ages 6—9) (ages 6-9) "Me" in sibship* (A) "Me" lst, sibs by parents (ages lO-lA) Boy "me" next mother* (2) (ages lO-lA) Independence Counterdependence Boy "me" next father* (1) ”Me" last, sibs/parents* Girl "me" next mother* (1) (3) (ages lO—lA) (ages lO-lA) "Me" isolated "Me" independent "Me" oppositional *Modified placements 62 Even with the above changes, the middle class still tend to cluster in the High Growth, the lower class in the Low Growth column of the Scoring Matrix. This tendency for the lower class to be less growth inclined is also manifest in the disturbed sample, which is 62 per cent lower class. This is in keeping with sociolo- gical studies which have demonstrated a positive rela— tionship between class level and mental health generally, and a correlation between lower class membership and schizophrenia in particular, (Gruenberg, 1961). Of special note is that with the proposed modifications to the Position sort scoring system, not only are all the hypotheses supported, but the disturbed sample literally is reversed in terms of the frequencies falling in the High Growth vs. Low Growth columns of the Matrix. The result is that the power of the Position sort to discriminate disturbed drawings is greatly increased. In addition, the agreement (in %) between the Position and Relational sorts rises from 3A per«cent to 53 per cent. The proposed modifications appear to be valid, in the sense that they have positive influence on several disparate factors. However, because of the _§ ppsp £3939 nature of the modifications, a new sample of draw- ings would have to be analyzed to test their validity. 63 Position Analysis--"Me" Placement The most significant additional findings in the H Position analysis concern the me" placement in the family group. The predominant placement is next to H H mother, followed by me next to father. Table 18 gives these placements for boys and girls, normal and disturbed. TABLE l8.-—Per cent Placement of "Me" Next to Mother and Father, Normal and Disturbed Boys and Girls. Group Next Mother Next Father Normal boys 38% 20% Disturbed boys Al% 2A% Normal girls 35% 20% Disturbed girls 22% 17% Of note is that 67 per cent of the disturbed boys with H H me next to mother are age ten and older, as compared to A3 per cent of normal boys so placed. These older disturbed boys account for the eighteen disturbed draw- ings shifted from the High to Low Growth column of the Matrix as discussed previously in the Position sort modifications. Although Table 18 shows no large differ— ences between normal and disturbed children in "me” placement next to parents, there are qualitative differ— ences in these drawings which will be discussed later, (see page 65). 6A Figure 7 (p. 52) shows the normal developmental trend in placements next to mother or father, and indi- cates that boys start out predominantly next to the mother in the lower grades, with placement next to the father taking over between grades V and VI. This is suggestive of one of the roles a father plays in helping free his son from dependency ties with the mother, enabling him to form an adult relationship with his father. It should be noted that the older disturbed H boys (ages 13 and 1A) did not have any "me placements next to father, and this at a critical period for them in terms of adult identity growth. For girls, the pattern in Figure 7 is one of early positioning by the mother which decreases with age. Placement by the father gradually comes to the fore, pre- dominates at grade V, but then gives way to the mother in grade VI. Although this particular configuration could be due to chance, there is also a logical explana- tion in terms of the father's role in helping his daugh- ter move away from the old, more dependent relationship with the mother in order that a new, more independent and adult one might flourish. Again, for the disturbed girls, there is no placement next to the mother at age 10 and above, suggesting a disruption of the mother- daughter relationship for some of these children. 65 For the disturbed boys and girls who drew the "me" next to one of the parents, there are signs in the draw— ings of a highly ambivalent parent-child relationship. Of the sixteen disturbed boys who drew the "me" next to the father, nine represented the "me" as big or bigger in body and/or head than the father. Of the remaining seven, five drew the father "gimped" or "goofy" in appearance (see below), while one showed a tiny boy over- whelmed by a giant of a father. These characteristics were extremely rare in normal drawings. Following are reproductions of 9 year—old boys drawings of father. They are normal, "goofy" and "gimped," respectively. One of the interesting things about these drawings is the eyes. Figure 8. Examples of normal, "goofy" and "gimped" fathers drawn by 9-year old boys. 66 Of the four disturbed girl's drawings with "me" next to mother, two suggested encroachment and one distancing. There was one head bigger than mother's, one overwhelm- ing mother, and two "goofy" looking mothers. All three disturbed girls next to the father portrayed the "me" as large as he, and the l2-year old was also heavily made up and holding his hand. These findings suggest that for these disturbed children, there is an approach vec— tor directed at a parent, but with avoidant vectors which create disruptive ambivalence. The major negative factors for boys appear to be father's impotence (physi— cally, mentally or emotionally), and mother's infantiliz— ing (from the large number of older disturbed boys placing the "me" by mother). For girls it might be father's impotence and mother's overbearingness or domination-—but without real emotional closeness. This suggests problems with fathers who have trouble feeling their own masculine potency, and mothers their femininity and maternalness. In most of the disturbed children's drawings with the "me" next to one of the parents, then, there appears to be strong ambivalence which causes problems with closeness, caring and respect. This is the best single indicator of emotional disturbance in the family draw— ings. However, it is hard to say if the problem is really one of impotence or overwhelmingness, or whether 67 it goes beyond this into a chronic lack of deep emotional contact in the relationship. If this is so, then with the advent of early adolescence and the confrontation with adulthood, the child may feel the last chance for growth is slipping away-—and the parental shortcoming most difficult to restitute also becomes the one hardest to forgive. Relational Sort Analysis For the Relational analysis, all the Hypotheses but IIB (younger vs. older children will tend toward High vs. Low Group Identity, respectively) were supported overall, but not in all demographic groupings. These were the youngest and the black children under Hypothesis IIA (disturbed will cluster in the Low Growth column of the Matrix), and the black children under Hypothesis IIC (disturbed clustering in the Low Growth column will not decrease with increased age). In considering the lack of support for Hypothesis IIA and the black group, Table 1A (p. A9) reveals that of a total of eight disturbed black children, two were placed in the High Growth column of the Matrix, whereas six were sorted into the Low Growth side, which is in the predicted direction. However, this difference was not large enough to overcome the effect of the 50/50 split of the normal black children between these two columns. Most of the normal effect is due to the disproportionately 68 high number of black drawings being sorted into the counterdependent quadrant; twelve out of a total black N of A0. This quadrant claimed 30 per cent of the normal black samples as compared to l6 per cent of the normal lower class sample. As was found in the Position sort, the normal black children tend toward counterde— pendence and isolation. The issue here seems to be whether this kind of isolation in the family group is really a condition of low growth for these black child- ren, or if it represents a kind of interpersonal adapta- tion which allows for growth. The fact that the dis- turbed black children tend more toward overdependence (63%) rather than counterdependence (13%) supports the latter idea. The need for the black child (especially the boy) to break away from the mother, and often without the help of a father, may be associated with this counterdependent loading. As in the Position sort, an unusually high number of the youngest normals appear as overdependent (quadrant 3). Chi—squares for 6, 7 and 8 year-olds were 6.81, 5.58 and 5.97 respectively. Of the total normal sample, A0 per cent of the 6 year-olds were sorted into this quadrant. In sorting these normal drawings, it was felt that they showed the same general characteristics as the other drawings in this quadrant; that is, the "group gimp" or "mashy" quality of the people in the drawing. 69 There also seemed to be a good deal of autistic embell— ishment, which led to applying the term "autistic license" to the normal 6 year-old drawing style. It would appear that the High-Low Growth dimensions of the Relational scoring system cannot be applied to younger children so generally. This is due to their far greater tendency to represent the overdependent, borderline type of group encroachment, or "mash" which gives the feeling of little breathing room in the drawings-~and which is so charac— teristic of many older, disturbed children's drawings. In looking at these drawings of the youngest nor— mals, it is difficult to tell if there are problems of incipient emotional disturbance, or if many children at this age are still caught in the web of overdependence. Perhaps the factors of dependence vs. overdependence as used in the scoring systems do not have real meaning at these earlier ages. Perhaps all of these things apply. It is also interesting to note that many drawings of older, disturbed children were so highly similar to these younger, normal children, who are perhaps emotionally "younger" than their age peers. This resemblance suggests that some of the disturbed children may be functioning at the level of chronologically younger children who may themselves be emotionally immature. The lack of support for Hypothesis IIB (younger children should cluster in the High Group Identity row, 70 older ones in the Low Group Identity row) for boys stems primarily from too many older disturbed placed in depen— dence, and too many younger normals in counterdependence quadrants. The older disturbed drawings in dependent quadrant 1 actually resembled to a striking degree the younger normals placed there. This raises the question of whether these disturbed children may be extremely immature rather than manifesting pathological symptom formation. They would perhaps be more in line with the immature character disorder as opposed to the neurotic or psychotic immaturity-plus-symptomatology. In quadrant A (counterdependence) the normals show a reversal, with the younger outnumbering the older children three to one. This is hard to rationalize in terms other than those applied to the black children's drawings in quadrant A with the "black factor”; that of counterdependence and isolation. Perhaps for younger children who feel over- whelmed by the family's demands on their autonomy isola- tion can be a technique for interpersonal adaptation and survival which is an alternative to emotional disturbance. For the girls, the lack of support for Hypothesis IIB is with the disturbed in overdependence and the nor— mals in the counterdependence quadrants. Disturbed girls tend to cluster in overdependence, but do not differ when comparing younger with older. The fact that girls play a more passive and dependent role in the culture when 71 older as well as younger may have something to do with the lack of age differentiation for the overdependence in the disturbed. The same effect for the normal girls as for the normal boys is seen in the counterdependent quadrant A, which offers support for the notion that counterdepen- dence may be a way of achieving pseudoindependence or freedom from overdependency. Hypothesis IIC (placement of disturbed drawings in the Low Growth column will not decrease with increasing age) was supported for all groups other than the black children. There were only four older disturbed black children in the total sample, and they did not show a reversal of the hypothesized direction, so at best the results are equivocal. More generally, however, there seems to be a problem with the black children in this sorting technique which suggests the possibility of a family/individual identity picture differing from whites, with resulting differences in the expressions of depen— dence, overdependence and counterdependence for them. As mentioned before, the black child probably has difficulty with his mother's protective and controlling ties on him, but there may also be a problem in how the black person must adapt to his environment generally. In striving for autonomy and "self," perhaps he tends to remain "hidden" and personally encapsulated, not only from the dominant culture but also within his family and subculture. 72 Counterdependent or isolation techniques would then become a life style. In looking at general effects of the sex and class variables among normals for the Relational sort, Appendix I, Table A0 shows that girls and middle class predominate on the High Growth side of the Matrix, boys and lower class on the Low Growth side. Breaking this down by quad- rant, there is the suggestion that dependence in the family is most common among girls, especially middle class girls, whereas independence is more characteristic of the middle class generally. Overdependence is more a lower class characteristic, whereas counterdependence is used more by boys; especially lower class and black boys. The predominance of middle class in the High Growth column and lower class in the Low Growth column is in agreement with the Position sort, and as mentioned before, fits with other data. The reason for girls predominating in the High Growth column comes largely out of their high fre— quency in quadrant l, but is also suggestive of important factors other than growth which are involved in the Relational sorting technique. One might be that girls are trained to be dependent, whereas boys are pushed more towards independence--which then leads to more counterdependent behavior in the family in lieu of true independence. However, this counterdependence in the family group would not of necessity be associated with Low Growth--except if it generalized to extra-familial 73 relationships. It would have been interesting to have drawings of a "group of children" to look into this kind of generalization. Perhaps the disturbed children in quadrant A (counterdependence) would also tend to show isolation in their peer relationships, whereas the nor- mals in that quadrant would not. Another interesting effect in quadrant A is seen in Appendix I, Table AA, where there are no significant differences between normal and disturbed for any of the variables, while almost all the other normal/disturbed differentials were significant in the other two quadrants. Of further interest is that inter-judge agreement on the normals sorted into quadrant A was the lowest of any of the quadrants (55%), whereas it was highest for the dis- turbed (70%). Perhaps the normals in this quadrant con— sist of two groups; those who are relatively normal, and those who have emotional problems that have not been identified because of the child's hiddenness. The argument could be raised that quadrants 3 and A are really not much different in content, anyway, because the judges might have tended to sort "sick- looking" drawings into the Low Growth half of the Matrix without really discriminating between overdependence and counterdependence. If this were the case, it would be clear that the frequencies in quadrants 3 and A would be roughly equal, as would the interjudge agreement. Table 19 indicates this is not the case. 7A TABLE l9.--Observed and Expected (50%) Frequencies of Family Drawings, Normal and Disturbed, quadrants 3 and A. Adjustment Quadrant O E x2 df p a 1;: 132:2 1 Normal 3 51 I; 6.18 1 .05 Disturbed E 3% 32:2 5,2u 1 .05 The interjudge agreement for quadrant 3 was 63 per cent for the normals, 65 per cent for the disturbed; for quadrant A, 55 per cent for the normals and 70 per cent for the disturbed. These data support the idea that there was a discriminating principle in the sorting of drawings into these two quadrants other than that they were just "sick-looking." Of further interest is the fact that of the 112 drawings that could not be scored, 67 omitted the "me," which would qualify them for placement in quadrant A above any other. Of these 67 drawings, 16 were done by disturbed children. However, in taking the 17 drawings of children age 12 and over, we find 12 of the 16 dis— turbed and only 5 of the 55 normals in that age group (x2 = 3A.lO). Not only does this make the omission of the "me" the best single indicator of disturbance in 75 children age 12 and over, it also is suggestive of some- thing about isolation or counterdependence in the family group in general, as was suggested earlier. In Table 19 (p. 23) it is seen that normal boys had 23 younger and 11 older members in quadrant A, whereas disturbed boys had 3 younger and 1A older. Apparently the counterdepen— dent or isolation technique changes meaning with age, for boys. For some younger boys, it is perhaps the only way they can function in the family group and retain some vestiges of "selfness." However, for older boys it seems to be more definitely associated with emotional disturb- ance rather than positive adaptation. Perhaps what began as a way of adjustment, with time and increasing age, became the road to increasing difference and "out of jointness" with others--with ultimate emotional diffi- culties. Uses for the Family Drawinggand the Sorting Techniques The results of this study suggest research, clinical and screening uses for the sorting techniques. Research uses would include looking at children developmentally, and in terms of the quest for individual identity, the struggle against dependency, the roles of parents and siblings, and the family dynamics generally. Perhaps an increased understanding of family factors in emotional disturbance might also be gained. 76 As a clinical tool, the family drawing reveals things about the individual within the family group that other tests may not, such as the child's negative feel- ings about himself and family members which are often painfully clear. Dimensions such as overcloseness or isolation can also be obvious. The final consideration is for using the DAF as a device for rapid screening of large groups of children to discover those who are handicapped in terms of their emotional growth. For this purpose, the Relational sort is impractical, as it is difficult to learn, and time- consuming to use. On the other hand, the Position sort is relatively objective, easy to learn, and takes little time to use. With the previously outlined objective modifications, and applying the qualitative indicators to drawings with the "me" next to a parent, the techni— que shows promise as a screening device. APPENDICES 77 APPENDIX A RELATIONAL SORT SCORING SYSTEM 78 79 Relational Sort Scoring System Quadrant 1 Real, non-demeaned people: Attempt to draw real peo- ple in real scene-—body, features, alive posture. Positive embellishments: Additions to people are enhancing of appearance. Fair to good body proportion: Overall size, proportions of head, limbs, etc. Closeness w/o encroachment: Perhaps affectional type contact, as of hands. Fair to good sex differen- tiation. Few omissions of major body parts. Low to medium stereotopy: May be some, but still see some differences in family members (hair, expression). Quadrant 2 Real, non—demeaned people, positive features: As above. Good to excellent propor- tions. No major omissions. Low stereotopy. Optimum distance between family members, may be doing own "thing." Quadrant 3 Unreal, demeaned people: Perhaps clown-like, mon- sters, robots; unreal features, expression, frozen position. Negative embellishments: Additions not enhancing. Poor proportions: Either overall size or body parts. Encroachment, overcloseness: Collisions, transparencies between family members. Poor sex differentiation among family members. Major omissions of body parts. High stereotopy among family members: Tendency to look similar in face, hair, posture. Quadrant A Unreal, demeaned people, negative features: As above. Poor proportions of body parts. I! H Major omissions in me. High stereotopy of group, not "me." Isolation of "me": either by distance, barriers, lower differ— entiation, smallness. APPENDIX B INSTRUCTIONS TO TEACHERS l. Administering DAF to regular classrooms 2. Administering DAF to disturbed rooms 80 A. 01-th INSTRUCTIONS FOR ADMINISTERING THE FAMILY DRAWING TEST Purpose Secure pencil drawings of a family from 600 students, grades 1-6, and from special classrooms for the maladjusted, in order to provide norms for the test. Materials 1. Paper (will be provided) 2. Pencil 3. Time; 15 minutes Directions 1. Distribute paper. Have students take out pencil only; (no crayons). Have students place name, age and grade in space provided on paper. Ask children to draw a picture of g_family; (Do not say "your family.") Let children know they will have about 10—15 minutes to.complete drawing. 6. Questions about how or what to draw should be answered by such things as, "That's up to you", or “However you want to do it.“ (#4 may be repeated, however 7. After 10 minutes, let students know that they will have only a few more minutes. 8. After 14 minutes, have students stop. 9. Ask children to place an "F" over the father, an “M“ over the mother, and "Me" over the person that most reminds the child of himself. .10. Collect papers. 11. Place a red check mark in the upper right hand corner of papers of Black childre 12. Place a blue check in the upper right hand corner of papers of students you-feel to be most improved in terms of their emotional growth. 13. Insert drawings in stamped, addressed envelOpes provided, and place in mail. Results of this study will be available to any of the educators involved in the study. 82 APPENDIX C COMPOSITION OF SAMPLE 83 8A TABLE 20.--Normal Sample: Boys and Girls by Age and Grade. Grade Boys Girls Both Age Boys Girls Both I 51 A6 97 6 17 19 36 II 36 39 75 7 35 A6 81 III 36 39 75 8 A8 38 86 IV 32 38 7O 9 26 39 65 V 23 37 60 10 31 32 63 VI 35 AA 79 ll 31 A5 76 -———— ———— 12 23 21 AA 213 2A3 A56 13 2 3 ‘5 213 2A3 A56 TABLE 21.-—Emotiona11y Disturbed Sample: Boystirls, Age, Source. Disturbed Classes Teacher Rated Both Age Boys Girls Both Boys Girls Both Boys Girls Both 7 2 2 A A l 5 6 3 9 8 A 2 6 2 2 A 6 A 10 9 A l 5 8 2 10 I2 3 15 10 A l 5 5 1 6 9 2 ll 11 A 2 6 8 O 8 l2 2 1A 12 5 2 7 6 l 7 ll 3 1A 13 A 1 5 A 7 A 8 l 9 1A 2 O 2 O O O 2 O 2 29 11 A0 37 7 AA 66 18 8A APPENDIX D BREAKDOWN OF POSITION SORT 1. By Grade: Normal 2. By Age: Normal and Disturbed 85 86 TABLE 22.——Normal Position Sort: Grade and Sex. Quadrant 1 By opp-sex parent lst; sibs by pars. Both positions Grade Boy Girl Both Boy Girl Both Boy Girl Both I # l9 9 28 2 2 A 21 ll 32 %(38) (20) (29) < A) ( A) ( A) (A2) (2A) (33) II # 10 9 l9 3 1 A 13 10 23 %(29) (23) (26) ( 9) ( 3) ( 5) (37) (26) (31) III # 16 A 20 1 2 3 17 6 23 %(A2) (10) (26) ( 3) ( 5) ( A) (A5) (15) (30) IV # 1A 7 21 0 0 0 1A 7 21 %(AA) (18) (30) ( O) ( O) ( 0) (AA) (18) (30) V # 10 12 22 0 A A 10 16 26 %