358 Journal of Personality Assessment, 1978. 42,4 Toward an objective Evaluation Procedure of the Kinetic Famlly Drawings(KFD) DAVID V. MYERS Universitv of Alabama Kinetic Family Drawings(FD)was examined. A quantitative scoring procedure was de- veloped from the clinical hypotheses Burns and Kaufman( 1970. 1972)to score 2l measurable KFD styles, actions, and characteristics. The scoring procedure was employed to evaluate sul:s indicated that four of seven sets of extracted component scores significantly differed veen ihe emotionally well-adjusted and the emotionally disturbed groups. One set of com ponent scores significantly differed between the younger and the oider groups, whrie twosets of scores did not differ among any of the four groups. The KFD iota score was found to dit nificantly only between the young emotionaly disturbed and the young adjusted groups, It was concluded that a quantitat: ve scoring procedure for Recently a promising new projective tion, and are found in actions associate drawing technique was introduced by with certain objects such as baiis. burning Burns and Kaufman(1970, 1972). view. fires, lights, etc. Styles, a third dimension ing emotionai problems in chiidren as interpretation, reiers to certain ap- stemming from the disturbances in inter- proaches to the drawing of K FDs that are personal, generally family, relationships. considered indicative of defensiveness these authors have deveioped a procedure and emotional disturbance. Compart known as the Kinetic Family Drawings mentalization (intentionally separating ( KFD) which provides a basis for exam family members by the use of lining) and ining the interpersona! dynamics of the Edged Placement(rectangularly placing hild. After 12 years of carefully studying all of the family figures on the perimeter of lyzing the page)ar linical significance of certain character- final aspect of interpretative significance tics. actions styles and symbols found in Burns and Kaufman 's KFDs, follow k ka fhaanterszc are ording to ities speptiat c Hasee n ghe ruano fa are eass commonly found in the drawings. Pencil ings to KFd symbois than they are in the erasures (reflecting ambivalence) and characteristics, actions, and styles hyper-extended arms Er Unlike most previous projective draw- to control the environment) are typical ings, the Kfd provides information that xampies of KFD characteristics. Ac- can permit the unique and valuable in- tions in KFDs are movements of energy vestigation of the personal and interper between people and objects. According sona perspectives of thechiid. However, Burns and Kaufman, certain objects the KFD, while promising, has two gen and depicted actions are said to ha era! shortcomings nergy invested in them. These"fields of First, Burns and Kaufman(1970, 1972) force"refiect such things as inhibition. offer no empirical validity for their hy anger,competition, and need for affec. potheses, but rather rely on the presenta tion of clinica cases which have been sub- d. Only two iatric Services for Children. richmond. Vir. studies have been reported with the KFD ginia, May, 976 O'Brian and Patton(974)attempted to
DAⅤIDV. MYERS Tabe Distribution of Groups of Subjects by Emotiona! Adjustment and Age Emotional Adjustmen: Disturbed Young( 6-8 years) YEA Oider(12-14 vears) OEA OED develop an objective scoring system via 4 Because of Burns and Kauf mans conten step-wise regression analysis. However. tion that the interpretation of KFDsym- their study was conducted prior to the re- bois is highly variabie. no attempt was lease of Burns and Kaufman's (1972)Ac- made to include symbois in the quantifi Id Symbois in Kinetic- cati Famihy-Drawings(KFD): An interpre he scoring procedure was employe rive Manual, and thus did not examine to evaluate K Fds obtained from two all of the variables considered significant clinically different groups of chiidren:o by Burns and Kaufman. Aiso, o'Brian determine the ability of the procedure to and Patton did not include emotionally differentiate between the two groups disturbed children in their sampie, fur. The clinical groups participating in the ther reducing the utility of their study. study were boys judged to be either emo McPhee and wegner(1976)exa tional well adjusted oremotionally dis differences in k fd styles betwee urbed. Further. the K FDs of each group but failed to specifically control for the ranges within each group IO CAetano mal"and emotionally disturbed cl examined were obtained from two separate ag ffects of age, intelligence, and other sensitivity of the procedure to age d quantifiable aspects of the KFD such as ferences. That is, each group was divided characteristics and actions. A second into younger and aidier subsamples yield shortcoming of the KFD technique of ing four groups of subjects (i.e, young Burns and Kaufman( 1972) is that no emotionally well ad usted [YEA]. young normative data are given with respect to emotionally disturbed [Y ED]. older emo- any developmental differences in chil- tionally well adjusted [OEA]. older emo- dren. in spite of the overwheiming evi- tionally disturbed [OED). These groups dence that graphic abilities are directly are presented in Tabie I related to age and intelligence(Good enough. 1926: Koppitz, 968 method The purpose of the present study was Subjects and materials to test the feasibility of employing a quan titative scoring procedure for evaluating ent study were i 16 boys, ages through procedure, the issues of both normative as lathe y dis yo ed. For each pro- be explored. For these purposes, an em- See Ap pibil o p 100:. Remi inad ance fora pirically-based KFD Scoring Guide was Microm. New however, here is to m ichang of ss oi developed by the present author using 21 Staps and styles derived from the hypotheses ot cations Phot Ic scoring procedures of Mc Phee (1974). a pho: o:0p anss ta: a cAnada, postage s $. o( 1
An Obiective Evaluation of the KFD spective subject, scores were obtained or ple of KFDs The average agreement be Peabody Picture Vocabulary Test tween two trained scorers was 94 (PPVT)(Dunn, 1965)and the Missouri a range of 81%0-100% Childrens Behavior Check List(MCBC) he or gina scores obtained using the (Sines, Pauker, Sines, Owen, 1969). KFD Scoring Guide with the 116 KFD Based on their ages and normal IQ scores protocols was transformed into a 20X20 on the PPVT, as well as their performance "within groups"prod n the MCBC. the boys were selected relation mat membership in one of the four groups. the four groups. One variabie Foiding Normal IQs were defined as faling within Compartmentalization, was deieted from the range of 80 to 12 the study since no subiect included that The 60 well-adjusted boys were selected style in his KFD protocol. To obtain a com a public school system in Northeast smalier and experimentally more man- For inclusion in the Ea grou each boy was referred by his teacher as components analysis of t he intercorreia- ng emotionally wei! adjusted, scored tion matrix was performed. Three criteria within the norma IQ range on the PPVT, were considered jointly in determining obtained norma to well-adjusted scores the number of components to be extracted on the subscales of Aggression. Inhibi- (a)the scree test. (b) the number of com- tion. Activity Levei and Sociability on the ponents yielding Eigenvalues greater MCBC, and had never received, nor been than unity, and (c)the component solu referred for specia! services for emotional tion providing the most easily interpre o2 learning probiems. ta ble structure for use in subsequent an- The 56 emotionally disturbed boys alves obtained from four units of the The results of jointly considering the gia Psychoeducation! System, a above criteria yielded a component so wide program for emotionally dis- lution that was employed to generate urbed children which accepts for treat- component scores for each component ment those referrals diagnosed as emo- The component scores were then ana tionaly disturbed by a Board-certified lyzed for each extracted component via svchiatrist. For inclusion: in this ED 2X 2 parametric factorial ANOVA. KFD up each boy, in addition to the psychi- total scores were also analvzed via a 2X2 atric diagnosis of emotional disturbance, parametric factorial ANOVA. In each scored within the normai IQ range on the case, the independent variabies in each PPVT and received scores within one ANoVA consisted of two levels of tandard deviation of the mean for psy- and two levels of emotiona! adjustmen chiatric popuiations on at least two of the In the ANova procedures, the 01 leve above four subscaies of the MCBC of probability was used as a criterion of Procedure sIgnificance Each subject was requested to pro- Restate duce a kFD according to the specifica The data obtained from the"within tions of Burns and Kaufman(1972. p groups intercorrelation matrix of the The testing situation was terminated scoring variables on the KFD were re when the subject ind cated that he was duced by a principai-components pro- finished and when the investigator had cedure. Seven components, accounting ecorded the childs description of the for 53%c of the total variance, were ex- drawing and had labelled the characters tracted and rotated via the varimax met in the drawing according to the child s di- od. The component ioadings of the 20 re each drawing so obtained was then ponents are presented in Table"en com- KFD scoring variables on the sey scored according to the k FD scoring ven components were Guide. Interscorer reliabilities of the 21 obtained for biect using the formu- scoring variabies in the guide were com- la given by For each puted on a random 27% of the totai sam- component. a 2 X 2 ANOVA procedure
DAVIDV MYERS Table 2 stated Component Loa dings for the Kinetie Family drawings (KFD) Components Phvs:ca: proximity m06000 0 Barrers 0 Relative Height ic 6. Arm Ex 8 7. Des 9 3 8. Safeiy of F:re 6-08-36 -12 9, Boy pa 9-8 ! I0 Rotors :2. Compartment 13. Underlining Figures 0 20 4. Bottom: lining -03 18 :6. Encapsulation OS :7. Edged placee:! 74 I&. evasions 02 41 0. Back placemen Eigenvalue .90}1s8 was performed empio\ ing the appropri- on Component were Shading Bottom te set of component scores as the de ining and Top Lin ng For Component pendent variabie. For each significant II scores. no significant main or inter- ifference obtained in each of the anais ses action effects were obtained. The var. o: the component scores. the variabies abies loading most high y on Component ontributing most highiy to the differ- !I were Force Fieids. Description of Ac d the highest component tion. Bodv Parts. Underlining of Indiv'ic loadings for the component from which ua Figures. Evasions. and Back Place alvsis o: varance of the Component i cant efe scores indicated a significant main effect ment, F(L. 12=46.62 p<, 0. but no for adjustment. F1, 112) =7.97. p<,0i. significant effects were obtained for age but no significant effects were obtained or the interaction. The varia bles loadin for the interaction of age and ad ustment. most h: ghly on Component!I were Bar Fhe variabies ioading most high v(+30) ers. Description of Action. Safety cf
Figures, Encapsulation, and Evasions. turbed groups. while Component V was Similariy, a significant main effect for associated with significant differences the Component IV scores were obtained between the young and older groups for adjustment, F(. 12=21616p<. 01. The KFD tota scores significantly dif- but no significant effects were obtained ferentiated the young emotionally weil for age or interaction. The variabies ioad- ad; usted from the young emotionaily ng most highly on Component Iv were disturbed Physica: Proximity, Description of A tion, and Number of Household Mem- The resuits generally support the feas- bers. On the Component V scores, a sig- ibiliy of employing a quantitative scor- nificant main effect was obtained for age. ing procedure with the kFD to differen- cant effects were obtained for adiust- tiate emotionaly wel: adjusted from Safety of Figures, Shading. and Com- with those component scores hold prom effect was found from the analysis of var. ise as discriminators of emotional ad lance of the Component VI scores for justment. Specifically, the results indi adjustment, F(I. 1 2)=17.01 p<oi cated that in inear combination with other variabies 1I variabies. (i. e. Pbvsi- but no significant effects were obtained ca! Proximity. Barriers, Description of for age or the interaction of age or adjust- Action. Body Parts. Rotations,Bo:tom Component VI were Body Parts, R Lin:ng, op Lining, Encapsulation, Evasions, and Num- tat:on, and Edged Placement. For the ber of household Members)differen main: or interaction effects were obtained. the emotionaly disturbed boys indirec- Componen: VIl were Reiative Height. tions that are consistent with the hypsman Erasures, Description of Action, and eses developed by Burns and Kauf Number of Household Members (1972) The resuits aiso indicated that the Likewise, the total scores of the kfd quantitative scoring procedure, was, in rotocois were analyzed via a 2x 2 part, sensitive to age differences bet weer NOVA. In that analysis, a significant the youngand oider groups. The anaiysis effect was obtained for the interaction of of variance of the scores derived from ge and adjustment 12)=7. 40, Component V resulted in significant ef <.01, along with a significant main fects for age, with the older groups scor- ffect for adjustment, F(, 12)=17. 68 p4 0. No significant main effect was the five variables that ioaded highly on obtained for age. Examination ofthe cell Component V, three of the variables(i.e means indicated that the KFD total scores force Fields, Arm Extensions. and Com varied as a function of emotionai adiust ment in the young groups bi rtmentalization) oniy ly loaded signif older groups. The young emotional cantly on one component and therefore well-adjusted subjects obtained signil differentiated between only the age canty lower total scores than the older groups. The two other varia bies,(i.e afety of figures and Shading) also weli-adjusted or ei: her of t he ED groups. ioaded highly on Components I and Ill A summary of the results is displayed respectively, and these components dif n fable 3 Of the compenents with com- ferentiated between adjustmen:. Thus ponent scores that s:gnificantly differ- in the present study, Force Fields. Arm entiated among the four groups Co Extensions and Compartmentalization onentsi, I. I, and vi were associated were not found to differentiate between with significant differences between the the emotionally well adjusted and the ell-adrusted and emotionally dis- emotionaly disturbed groups. and there-
DAVIDV. MYERS Summary un: Component Analysis and Analysis of variance G Direction ariabic DiT::e b Variabie Loaded[ Cum D: ferene i Physical proximity Distent EA<ED B ArrEsten: Reiative height 4. force fields Yoar: g< older 6. Arm i xtensions Young nicer :, Deserint: on of Action: ELI! IV VIE Ac:ustment EA<ED s Satety o: Figures EA<ED I. Rotations FA<ED YOung <oer Toling old ! :4. Bottom lining Adi:stment EA<ED Adjustmen: 46. Eieaasutali t Juristen: EA<ED :?. ideea placement Adus:nien: EA<ED :8. Easi Li HIL.ViI EA<ED 冫9. Nimber ot meme;s Auiusizrent FA<ED C Back Placement fore shouid not be automatically accepted ferentia: information. since these variables as indicators of emot onal disturbance, aiso loaded on components that differ particularly when scored by the present entiated bet: the emotionally well scoring procedure. The presen: evidence ad:usted and the emotionally distur bed suggests :hat these three variables in com- groups. O: the remaining six variables bination with the other variables of Com- four: i.t. Relative Height. Erasures. tn ponent v, bear little reiationship to emo. declining Figures, and Back Placemen: tional ad ustment Safety of Figures and did not load highly on any of the compo Shading. the remaining varia bies ioac: nents which were associated with s:gn!: high:y on Component V provided no dif- cant di: ferences among the groups. Th
364 An Objective Evaluauion of the kFD KFD total scores signi! icantly diferen esent study,'s findings contradict those iated between the voung emotional cphee and wegner, it is possibie that well adiusted and the emotionally dis- the KFD sty ies appear more often in ED turbed but did not differentiate among groups when the styles are measured it the three remaining groups combination with other variables. Addi The findings that scores derived from tionally, it is possibie that differe ces IH certain components were sensitive to age the sampie characteristics could account sensitive to both age and ad ustment, to styles. That is, McPhee and wegner detracts from the utility of the present compared the KFD's styles with en quantitative scoring procedure. Also. tional"!"children with the KFD that the tota: scores were insensitive to styles of emotionaily disturbed chidren groups, further underscores the limited KFD styles to the emotionaly well-adjusted effect:veness of the scoring procedure children with those emotionally disturbed Finally, that the findings of the pres The findings of the present study are ent study indicate that scoring system generally consistent with the esearch on effecti does not suggest that the same system scoring procedures used ith human wil: provide reliable diagnostic informa figure drawings to differentiate among tion in the individ uai case. In fact. the clinically different groups ie, g, Reznikoff analysis of the means and standard devi tomiien. 1956). whiie the effective- ations of the component and total scores ness of the KFD quantitative scoring pro- in the present study indicate that the different groups does not provide incon trovertibie evidence for its validity it does the individuai case and should be used add support to the concurrent validity of only as a research too! from which more the KFD In the present study, clinica! sensitive scoring approaches can be de groups were defined by scores on the Mis. veloped souri Children,'s behavior Check List (S:nes, Pauker, Sines. Owen, :969)and Burns, R c.&K by psychiatric diagnosis. That the KFD Drawings(KFDF An intro scores derived from the cuantitative scor- GrEn15 1 ing procedure aiso effectively differenti- New york: Brunner: Mazel. 97G ated the ciinical groups is consistent with Burns. R. C.&Kaufman. S. H:. Acnons, stries other studies indicating that the kFd New York: Brunner Mazel agrees with concurrent psychometric and Dunn. L behavioral data. (e. g. O Brian& Patton 974:Sims,1974) The findings of the present study are however in contradiction with several the findings of McPhee and Wegner( 1976). Goodenough, F L. Measurement of intelligence in a study examining KFD styles, Mc an ings. Y inkers. Worid books Phee and wegner concluded that the six les defned by burns and Kaufman Koppitz, E M. Perchoiogical evaluaiion ofc 1972)occurred more frequently among stratton. 968 emotiona: !y norm a! chiidren. In the pres ent study, the styles of Bottom Lining. McPhee. 3. P. Empirice! evaluction of tine kine:ic Top lining, Encapsulation and Placement. in Finear combination with an the emotionally disturbed than in the McPhee. J P, wegner K w.Kinetic-Family emotionally weli-adjusted groups While it is not entirely clear why the ment:975.44,487-49
DAVDV MYERS OBran k P. &raton. w.F. Developmen ci ar David v Mvers. PhD swine, orne u! Fersonadt: ssess men,, The trivers:tv c Reznik ot f M. Tambien D The tse sf human i::y Aiabama 354%. Eure araw:ng it tne diagnasis s: crEatiE pa cri rna of ne Pyscho:eg. 24 Received sims. C.A. Kinetic Famiy Drawings and he Revised anuary 26.98 O. Pauker.F D. Sines. K. ower