G A T I N G 1‘ THE APPLICATION OF A PERCEPTUAI. THEORY TO THE ISSUE OF DIAGNOSIS IN PSYCHOTHEIAI‘Y ROBERT BERNARD LEVINSON ~ A THESIS FOR DEGREE OF PH 0 MICHIGAN STATE UNIVERSITY 1957 This is to certify that the thesis entitled THE AFPLICATIUN UF H PtRC.rTUAL THLoRY TO THE ISSbE OF DIAGNOSIS IN PSYCHUTHERAPY presented by ROBERT B. LEVINSON has been accepted towards fulfillment of the requirements for 41L degree in. PS YCHOL OGY ,‘ ., ~. xi f» ’A ‘- // x- / < /.. Q I I 1‘ I . ’2'”'*—‘ l‘ i ., v" . . ~ I] . . ‘1 .‘I A . Major professor Date “ARCH 7, 1958 LIBR A R Y Michigan State University G A T I N G : THE APPLICATION OF A PERCEPTUAL THEORY TO THE ISSUE OF DIAGNOSIS IN PSYCHOTHERAPY BY ROBERT BERNARD LEVINSON A THESIS Submitted to the School for Advanced Graduate Studies of fiichigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1957 "And once set up, a system probably does as much harm as it does good. It serves as a sort of sacred grating behind which each.novice is commanded to kneel in order that he may never see the real world, save through its interstices." Edward C. Tolman ACKNOWLEDGEMENTS In the course of pursuing a doctoral degree an individual becomes indebted to many people for the assistance offered to him. At this time, the author of this dissertation would like to acknowledge the help he has received. For the guidance in helping the author steer through the sometimes troubled waters of advanced Psychology courses and for serving as the original chairman of this doctoral committee, before going on a sabbatical leave, many thanks are due to Dr. Donald M. Johnson. To Dr. Donald L. Grummon who graciously accepted the role of chairman of this doctoral committee, and who gave generously of his time and effort to the development of this investigation, the author wishes to express his gratitude. The author would also like to express his apprecia- tion to the other members of his doctoral committee: Drs. M. Ray Denny, Bill L. Kell, G. Marion Kinget, and Louis L. McQuittye A debt of gratitude is owed to the Veterans Admini- stration for making it possible for the author to conduct this research by providing not only many of the clinicians who served as Judges, but also by making time available for the collection of data. This study could not have been conducted without the cOOperation of the clinicians who served as Judges and Mr. ”L" who was the client. To the Judges who are listed in Appendix A and the client, Mr. "L", the author offers his many thanks. Finally, to the two women who made this whole project seem worthwhile,.this work is dedicated: "Thanks Betty." "Mindy, maybe you’ll get to know your Daddy now." vi otl api: Vi t1 9V9: AN'ABSTRACT Gating: The Application of aaPerceptual Theory to the Issue of Diagnosis in Psychotherapy by Robert Bernard Levinson Gating is a process introduced by Bruner (1) as part of a perceptual theory. It is a mechanism which permits the pas- sage of certain stimuli in the environment (those in accord with an individual's category system) and blocks passage of other stimuli (not in accord with the category system). A category system may be conceived of as a network of sets, or personal constructs, which evolve from a process of learning how to isolate, weigh, and use critical attributes or cues, and which is markedly influenced by individual expectations and needs. The process of gating has direct implications concerning the use of formal diagnostic procedures in psychotherapy. The commonly accepted view —- "diagnostic" position -- holds that pre- therapy plans (based on diagnostic techniques) enhance a ther- apist's understanding and, thereby, increase his effectiveness with a client. Adherents of an "a-diagnostic" viewpoint, how- ever, suggest that pre-therapy formulations may blunt the the V6 a» therapist's understanding of his client. That is, the oper- ation of a gating mechanism could result in a selective screening out of certain of the stimuli present in thera- peutic interviews--namely, those cues not in accord with the pre-therapy schema-whi1e, at the same time, other cues (more in accord with the pre—therapy plan) may be over-evaluated. This study concerned itself with the following general proposition: the type of preconceived categorization system with which one approaches events will affect the sensitivity to nuances in those events The proposition was investigated through two, more specific, hypotheses. Hypothesis 1: the sensitivity of an individual to nuances of a situation will be diminished if he ap- proaches that situation with a preconceived categorization system Hypothesis 2: the sensitivity of one individual for another person will increase as the degree of congrup ence between the categorization systems of the two individuals increases Using a Q-methodology design, 32 clinical psychologists, (in four groups of eight each) after doing a self-sort, at- tempted to predict a client's self-concept (as revealed by Q- sorts) on the basis of diagnostic and interview material. Each Judge made two predictions. The dependent variables were the accuracy scores-~the degree to which the predictions approximated the client's actual sortings. The independent variable was the sequence in which the stimulus material was presented. Four sequences were employed: (i) Interview (1") material before Diagnostic (D) material; (ii) D I"; (iii) 1' I"; (iv) I" I"--conditions (iii) and (iv) represent con- trols; I' and I" being, respectively, the 11th and 13th tape recorded therapy sessions. Using an analysis of variance technique, the results obtained indicate: Diagnostic material does not markedly enhance nor significantly decrease the sensitivity of a Judge to Interview material; and, similarly, for the effect of Interview material on Diag- nostic material Although the presentation sequence of the stimulus material does not significantly af- fect sensitivity, making two predictions, particularly if based on widely differing types of material, results in significantly less accuracy for the second prediction Support was found for the validity of Hypothesis 1 The validity of Hypothesis 2 remains in question; the results were in the anticipated direction but did not reach the level of statistical significance The general proposition underlying this study was, in the main, supported by the findings The limitations of this investigation, as well as its implications for clinical practice and future research, were discussed. (l) Bruner, J.S. "On perceptual readiness" Psych. 3&1. 1957 641123-152 TABLE OF CONTENTS Acknowledgements Table of Contents List of Tables List of Figures INTRODUCTION BACKGROUND OF THEORV AND RESEARCH Diagnosis and Psychotherapy Diagnostic Viewpoint A-diagnostic Viewpoint Perceptual Theory Theoretical Implications HYPOTHESES TO BE TESTED General Porposition Hypothesis 1 Rationale for Hypothesis 1 hypothesis 2 Rationale for Hypothesis 2 l~LETEiOD Overview Components of the Experimental Design Th Judges The Other The Input Diagnostic Protocols Interview Haterial The Outake O—sort Methodology The Q-arrays The Scores Interpretation of Scores Procedures Procedure for Other Procedure for Judges EXperimental Conditions Additional Control Features Experience Time Spent with Stimulus Material Fatigue Memory Interaction between Q-sorts Social Desirability of Items ii p. p. p. p. p. p. p. p. p. é p. p. p. p. ‘v\ 1..) O p. p. p. p. p. P- p. p. p. p. I). p. p. p. p. Do p. p. p. p. p. ii iv vi RESULTS AND DISCUSSION Reliability Other Judges Results "Robot'Accuracy Hypothesis 1 Hypothesis 2 Limitations GENERAL IMPLICATIONS AND SUGGESTED FURTHER RESEARCH SUMMARY AND CONCLUSIONS Bibliography Appendices List of Judges Equating Procedures for Social Desirability Suggested Instructions for Q-sort Statements of Q' and Q" Arrays Rating Scale Experience Data Sheet Method of Selection of Interview Material Rationale for Exclusion of AS Score Effect of Different Levels of Experience on ACC Analysis of Data: Scores Based on Items LcHtflQ’EJDjUO.Uj:I> iii LIST OF T'PLES Table I Required Frequency Distribution for p. 35 all Q-sorts Table II Range of Judges' Clinical Experience p. 39 under Four Experimental Con— dition Table III Reliability for "Other” p. #6 Table IV ’ACC and RS Correlations for all p. #9 Judges Table V Number of Judges Achieving ACC Scores p. 50 different from Vero at .05 and .01 Levels of Significance Table VI Number of Judges Achieving ACC Scores p. 51 Beyond Robot's Eiducial Limits Table VII Analysis of Variance of ACC Scores . p. SE for Four Groups of Judges under different Experimental Condition with Two Sessions for each Group Table VIII Analysis of Variance of ACC Scores p.56 Applied to I'I" and I"I" Con- ditions Table X Analysis of Variance for Counter~ p. 57 Balanced Tests Applied to I"D and D I" Conditions Table X Analysis of I"D Condition; Judges p.105 Dichotomizaion Basis of both Diagnostic and Therapeutic Experience Table X Analysis of D I" Condition; Judges p.106 Di hotomized on Basis of Therapeutic Experience Table XII Analysis of D I" Condition; Judges p.107 Dichotomized on Basis of Diagnostic Experience Table XIII Analysis of I'I" Condition; Judges p.108 Dichotomized on Basis of _ Therapeutic Experience iv v'TV 4L4. XVI I"JII XVIII 'r-rwr «(Lids anr JHL Analysis of I'I" Cond’tion; Judges Dichotomized on Dasis of Diagnostic Experier ce Analysis of I"I" Condition° Judges J I. , . l .’ a chnotomlzed on SaSIS OI Therapeutic Experience C4 Analysis of I"I" Condition; Dichotomized on Basis Diagnostic xperience O H) $3 C4 (N (D 01 Analysis 0: All Conditions; Judges Dichotomized on Basis of Therapeutic Experience Analysis of All Conditions; Judges Dichoto mized on Basis of Diagnostic Expe: ience ACC and RS Correlations for All Judges (Based on Items) Analysis of Variance of ACC Scores for Four Gro aps of Tudges udder '14 491‘ lllerent Experim eital Conditions each Gloup \‘Iiufl. TWO 385810115 .LOI' (Based on Items) ” 'r r1r" ‘; ‘1 "1T rii'Tfiq" ‘ ‘ | I ~> ‘ . I‘ ‘- .L \1 kl‘L—VHLJ r“ J.° ~.¢ . 4. . ucileln3.olC FI‘CSQHCC; ' on of procedures for all Judge 0 .1 ul- 1" Q 'r u" '1 J >I‘ ’l" x‘ * ll: )‘V ~' ~.r #4 I . t I . is u) INTRODUCTION It is the intention of this study to bring some empirical evidence to bear on the issue regarding the advisabilityof pre—therapy diagnostic formulations. In.view»of the amount of time Spent in doing diagnostic “workdups" for therapy candidates (cf.87,p.101-2) as well as the amount of heat generated in.discussions be- tween clinicians of differing persuasions regarding the usefulness of such reports, it.was felt that an;attempt to bring some light into this controversy would be'a worthwhile endeavor. Further, this study will also serve as an empirical test for an;hypothesis translated.from the area of per-. caption into the realm of clinical psychology;. Critics, as well as cliniciansthemselves; often bemoan the lack of well formulated theories in the clinical field. PerbA naps. studies involving such translations can lead to a rapproachment between the more theoretically oriented areas.1n:psychology*and.clinical.psyohology$ To be more eXplicit, this investigation is concerned with the general hypothesis that an individual's sensi- tivity to the nuances present in a situation is markedly affected by the type of preconceived categorization system with which he approaches that situation. That is, if, say, a clinician were to enter into a therapeutic relationship with pre-formed "notions" about the client (based on, say, diagnostic protocols) such a therapist might be less sensi- tive to certain cues in the therapy session than if he had not had this presumptive formulation. The focus of this investigation is the effect that preconceived ideas have on sensitivity. In what follows the theoretical background and em- pirical research pertinent to the above area of interGSt will be reviewed. The definitions and assumptions under- lying the general and specific hypotheses, as well as the rationale and procedures for testing these hypotheses, will be detailed. A discussion of the results obtained and their implications for future research will follow. The final chapter will contain a summary and the conclusions of this inveStigation. BACKGROUND OF THEORY AND RESEARCH Diagnosis and Psychotherapy Psychodiagnostic procedures appear to have earned a dual role in relation to psychotherapy. The first use made of diagnostic formulations is to provide a foundation for devising a plan (1,71,106) or establishing a rational (99) and scientific (89) basis for the choice of treatment (10). In its second role diagnostic techniques serve as adjunctive procedures in therapy. In this latter function they are employed to: direct therapeutic activity toward fundamental problem areas (62,72); establish rapport (5,21); overcome plateaus (62,72) or resistance to interpretations (5)-in the therapy sessions; as a means of becoming better acquainted with the dynamics of the client's thoughts (5); and, to sup- ply information for use in future interviews.1 In both these roles, moreover, there is a close inter- action between the two fields of clinical interest. In Watson's words (106,p.26) "If diagnosis does not imply treatment, the diagnostic study would be of little value." 1 A possible third role for diagnostic procedures might be delineated-~research. Muench (73), for example, cites the following ways in which diagnosis has been oriented to ther- apy: to determine treatability, to determine success, to study the process of therapy, and to make compatative eval— nations of therapeutic techniques. -3... Ty The most commonly accepted generalization made from this diagnostic-therapeutic interaction is that it has resulted in cross-fertilization of both fields. For ex- ample, the insights achieved in therapeutic interviews have lead to a verification and enrichment of diagnostic interpretations (e.g.,90). Despite the wide acceptance of this mutual-enhance- ment viewpoint, there is not universal agreement among clinicians that Such is actually the case. To some prac- ticing psychotherapists the advisability of making the results of diagnostic tests known to the therapist who will see the client is a point at issue. While the adherents of the more traditional "diagnostic" point of view contend: "The diagnosis...is an essential part of the approach to the case" (15); prOponents of the opposing or "a-diagnostic" viewpoint maintain: "Diagnostic knowledge and skill is not necessary for good therapy...." (88;cf.87,p.101-2). Thus there are two conflicting positions on this issue. The rationale for each of these points of view will be ex- amined more closely in the following sections. Diagnostic Viewpoint Diagnosis is conceived of as a formal description of a client and his behavior by a variety of methods whose basic purpose is to discover the personality dynamics of each individual case. It is implied that the more complete 4+- the description, the more complete will be the therapist's understanding of why and how the individual got that way. This, in turn, will enhance the therapist's effectiveness, and, hence, the effectiveness of the entire therapeutic undertaking: "Diagnostic constructs should sensitize the clinician to respond to significant characteristics of the client's behavior that might otherwise have been over- looked" (10). In support of this thesis Thorne (100) lists the fol- lowing as among the important objectives of diagnosis: 1. to demonstrate the aetiological factors 2. to differentiate between organic and functional disorders 3. to discover the personality reaction of the organism to its disability h. to discover the extent of organic change with resulting functional disability 5. to estimate the extensity or intensity of the morbid process in relation to actuarial data concerning type and severity 6. to determine the prognosis or probable course 7. to provide a rational basis for specific psychotherapy 8. to provide a rational basis for discus- sing the case with the patient and relatives 9. to provide a scientific basis for clas- sification and statistical analysis of data In Thorne's words: "...the eXperienced clinician utilizes diagnosis as the foundation for all rational case handling" (100). A-diagnostic Viewpoint Opposed to the traditional orientation, outlined above, are those clinicians who see diagnostic formulations as pos- sible obstacles to an effective therapeutic relationship. Perhaps the most vocal group holding this "a-diagnostic" view are those therapists of the Client-Centered persuasion. Rogers (88) writes: I The therapist must lay aside his preoccu- pation with diagnosis and his diagnostic shrewdness, must discard his tendency to make professional evaluations, must cease his endeavors to formulate an accurate prognosis, must give up the temptation subtly to guide the individual, and must concentrate on one purpose only; that of providing deep understanding, and accept- ance of the attitude consciously held at fins moment by the client as he explores step by step into the dangerous areas which he has been denying to consciousness. The rationale offered for this point of view cites two basic objections to the "diagnostic" approach: ...the very process of psychological diag- nosis places the locus of evaluation so definitely in the expert that it may in- crease any dependent tendencies in the client.... When the client perceives the locus of judgment and responsibility as clearly resting in the hands of the clini- cian, he is, in our judgment, further from therapeutic progress than when he came in. (8%,p.223) ...it has certain social and philosophical implications which need to be carefully considered and which, to the writer, are undesirable. ...the long-range social implications are in the direction of the social control of the many by the few. (8H,p.22h) -6- WIKEEIIIJ fl In addition Rogers maintains that diagnostic formu- lations may blunt rather than enhance the therapist’s understanding of what the client experiences during the therapy sessions. Other objections, based on empirical evidence and reievant to several of Thorne's points (p.5) can be cited. Powdermaker (80) notes that the diagnostic approach leads to "a gathering of data which are not obviously and im- mediately concerned with the patient's problem;" (cf. Thorne's point 3). Soskin (93) found both experienced and novice clin- icians, when using the Rorschach and/or the Thematic Ap- perception Test as a basis for conceptualizing a subject, appear to be predisposed to anticipate more maladjustive tendencies; (of. Thorne‘s point 5). Patterson (77) questions the usefulness of diagnoses in selecting a therapeutic approach "since the choice de- pends more on the specific training, experience, and preferences of the therapist than upon the diagnosis;" (of; Thorne's point 7). Further, Herrick & Nagy (56) found a correlation of .15 between the ability to make a diagnosis and the ability to choose prOper remedial techniques. Ash (3) found the reliability of psychiatric diagnoses to be .20. Mehlman's (70) more recent study concluded: "The existing system of psychiatric classification can probably -7- I ..i have little value for administrative management of patients, or for research...." (of. Thorne's point 9). Gill 23.§l.(49) refer to diagnostic procedures as char- acteristic of “the older psychiatric textbooks;" and, Coleman (16) contends: "Nothing of importance except the patient as such and the problem of his treatment is left out in the diagnostic approach." Thus are the lines clearly drawn which separate the op- posing views regarding the usefulness of diagnostic formu- lations in psychotherapy. Perceptual Theory An empirical method, capable of evaluating the validity of the above viewpoints, was suggested by Bruner's (11) recent theoretical paper on perception. His formulations serve as the theoretical framework for this investigation. In discussing perceptual readiness Bruner offers the thesis that perception involves an act of categorization.2 He conceptualizes the perceptual process as follows: Put in terms of the antecedent and sub- sequent conditions from.which we make out inferences, we stimulate an organism with some appropriate input and he reSponds by referring the input to some class of things or events. ...On the basis of certain de- fining or critical attributes in the input, 2 "To categorize is to render discriminably different things equivalent to group the objects and events and people around us into classes, and to reapond to them in terms of their class membership rather than their unique- ness." (12,p.1) -8... wast. are usually called cues.....there is a selective placing of. the input in one category of identity rather than an. other... ...A theory of perception...needs a mechanism: capable. of inference and cats— gorizing.... ...all perceptual experience is necessarily the end product of s- m.- gorizstion process. (.11 ,p.123) Tue reasons are offlered for this. position: first, the perceived event achieves its meaningtfrom a class of percepts with which it is grouped-wategorizing is: equivalent to giv- ing identity to a perceived. event. The categorisation system that an individual develops evolves from. a- process of learn- ing how. to isolate, weigh, and. use critical attributes or cues for grouping events in. equivalence classes. This view regarding the use. made of cues is similar to Brunsuk's probabilistic theory of perception. (1%) A second feature of perception is, what. is perceived is "varyingly veridicale-what is perceived is somehow a rep- resentation; of- the external. world. Bruner. suggests that veridieelity is not somuch a matter of represmtstien as it 18' "30d“ building” ((21.1 05,1).1 85-62).- In learning to perceive wears learning the relations that exist between properties of objects and events that we encounter, learn- ing apprOpriate categories and category systems, E1332 p1 MBMMEMM 1,. Bruuer contends. that a: fruitful way, of thinking about perceptual readiness is in terms. of the accessibility of categories for use in coding. or. identifying events in: the .9. O... swt ate IBtU «mic Mme: We: send Bnme flux: environment. the likelihood that a sensory; input till he categorised in terms of a given category‘is not cnl'ys letter at fit between; sensory! input end category specifi- eatim, it depends else on the accessibility, of a category. Lecessibility is defined in toms ct. measures: the greater the accessibility. at a category (a) the less the input necessary for categor- nation to occur in tom of this category, (13) the fider. the range of inpu ut chmcter» istics that fill be ”accepted as fitting the category in question! (e) the more likely that categoriesthat a- better or equally good fit for the” input fill be naked. (11,p.129) the relative accessibility of a. category ‘ depuds on. the meetatius of enindividuel regardingthelikelihod or events to be neeuntered in his. environ-mt, and, upon the search requireeents. imposed on. the individualby his. needs. Brener. sees- the categorisation-.prccess involved in percep- tionaas a. tour step decision. sequence. ‘1 P” .5?" °“ fmenmsesm that results in the perceptual isolatim od‘ an object or an “tun" 2) search-c-‘a second process. of more pre- c se cant based m additional cues which my be equally silent or 'unemc- scious' 't 3) Cm: --‘ on check-J‘s tentative Wt at Id'enfifim occurred, the search is narrowed for additional contimto genes to check this placmt. ...He ' » spea at a selective ' ting’ process coming into operation 8'this stage, hav- ing the etzect of. reducing the effective input of stimulation not relevmt to the confirmatory process.“ ~1c- . . ‘- 3"“. QC. - g... a 5mm to additional m. is either normalized or ”gated out' w (1.1, p.130-1) - Beginning in step three, Bruner' has. introduced the central theoretical concept for" tbs;- present MW. Gating: is seen as a: mechanism which permits passage or certain stimuli in the environment (those in accord fith. the subiect’s category? system) and blocks passage. of other stimuli (those not in accord fith the category system)... This. is one of four mechanisms which mediate perception and perceptual read- inessv-(i); grouping and integration; (ii) access ordering; (iii) match-mismatch signaling , md, (iv) gating. In support or his cmtenticn regarding the eristmoe od’ some type of gating. mechanism, Bruner cites recent physio» logical findings. related to neural feW..mes. mt is, not onlygdoes innervation. of a sensormreceptor lead to a muscle. response (as in the reflex arc response) but the activity of the muscle changes ,; to some. degree, the status a the receptor. Such .. a sequence. of events. can. result in the receptor. becoming sensitised or «fie-sensitised. In the torn- er instance it would result in an increased readiness to re- spmd-perceptual vigilance-om the latter instance the de- creased readiness to respond could be termed percepmai de- fense. Thus. a filtering or. gating system does not require the nechinati one of a. master-mind homunculus. Rather per. .11.. ceptual gnreadiness is- due to interfierence: ' ...ieilure to. perceive ..is most. sites not a lack of perceiving but a matter of inter- fame with perceiving. ....the. inter'fe—r-eEce More categorizations‘ in. highly access- ible categories that. serve to block alterna— tive categorisations in. low accessible cate- goriess [hm-.116 ct.(151) 3 Whom-a subject is: presented with a: red four of clubs its» chistoscopicelly (.13) the tendency is. to report a- four. out diaonds-or a four of clubs with- the color-suit relation:- ship rectified. That is, the. accessibility of. the category as red tour of clubs" is very. low, while the accessibility of the category ”a red four of diammdss'.’ or *a black four- af clubs" is much higher. Thereported. results indicate that the input is. categorized in terms. of. the. more accessible categories, despite the perceptual. distortion. that this may entail. It is this "degree of openess or closednessV-greater or lesser accessibility of categories. to. sensory..- input- vhichis mediated by the gating process. Stimuli approwiate for closed or low-accessible categories are "blocked" or "gated out,” as in. the experiment cited above, while stimu- li tor open or high-accessible categories are "gated in". the similarity between. "gating" and. the concepts. of; set and attention is striking. Set, which may. be defined as a readiness to make a. meciried. response to a. specified stimulus (57.11.65) ,. also has both: a positive and negative .1 2. aspect; to be set for.- one kind of. response. or one kind a stimulus necessitates. an inhibition. or inattention to other events in the _mvironment. The universality of the concept or set is suggested. by Allport: ...“being set“ for a. certain stimulus or for a reaction to it, and the oftect- which. this setting has upon the ensuing. or. corresponding behavior are phenomenathat are practically as. universal as behavior itself. (2,p.21-1) Johnson (57.12%?) points- out that sets' may be establish- ed by motives (80), social interaction (1+7), past. experience (92), or my be induced by the task the. subject is performing (66). It is suggested. that the accessibility of an individu- al‘s categorization system can be. ”gated" in. similar fashion. For Mple, Jones. a de Charms (58) report that different in- ferences will be drawn from the. same behavior as a tunctim or the set promoted by. the structure of the social situation. In view of these considerations, the following, deri- vations appear. warranted. Individuals approach. environment- al events with a. “built-in” categorization system or a net- work or sets; (Kelly (60.9.8) would speakin- terms of "trans- mrent templates'“ made or personal constructs). Such a system serves to filter out, to a greater or lesser extent—gating“ those aspects of the input stimulus which; are mt in accord with this. system. In other words, stimulus elements from an event which fit pro-existing categories will be quickly .13. ‘ . perceived--the prinCiple of prior entry3--whereas, other stimulus elements, not fitting the individual's categori- zation system as well, will be perceived less readily, and may even be filtered out completely; i.e., perceptual defense. There is some eXperimental work extant on perception which lends support to Bruner's position concerning the gating phenomenon. For example, Wyatt & Campbell (109) have shown that if a subject deve10ps a wrong hypothesis about the nature of a stimulus presented under sub-Optimal conditions, the perception of the stimulus in terms of its conventional identity is slowed down-~there is "a reduction inthe adequacy of perception due to previous unverified hypothesizing or guessing." Postman & Bruner (79) have also demonstrated the detrimental effects of premature hypothesizing on veridical perception. lheoretical Implications ' A formulation such as this has direct implications for the issue concerning the use of diagnostic procedures in therapy. 'When diagnostic considerations have lead to the development of a plan for therapy--i.e., the client's dynamic structure, as well as his major areas of conflict have been delineated through diagnostic procedures, and 3 When two stimuli occur simultaneously, the one to which the individual is prepared to attend is experienced first (2,p.215). -1h- strategies have been formulated to cope with these factors in order to attain certain specified goals-~in such a sit- uation, it may be said that the clinician approaches the therapy with a preconceived category system. The commonly accepted view is that this plan provides additional understanding and consequently increases the effectiveness of the therapist. However, as Bruner suggests is the case when dealing with perceptual stimuli, it may be that as a consequence of the diagnostic procedures cer- tain categories-~namely, those which are in accord with the plan for the therapy—~are more accessible to stimuli in the therapeutic sessions than are other categories, not in ac- cord with this schema. It is as though the therapist is selectively attuned to receive certain types of information and to screen out other types. In effect, the therapist has become less sensitive to the entire range of stimuli. Seen in this light, deductions from Bruner's theoretical position would suggest that therapy based on pro-conceived plans could make a therapist less effective, insofar as being aware of theintal range of nuances in therapy inter- views is considered a positive factor in such a relationship4. # The validity of this statement rests on the assumptions one is willing to make; cf. discussion in Limitations, p.61. -15.. HYPOTHESES TO BE TESTED This investigation is concerned with confirming or rejecting two major hypotheses which relate to the fol- lowing general prOposition: the type of preconceived categorization system with which one approaches events will affect the sensitivity to nuances in those events A categorization system will be termed "preconceived" when it is introduced into a new situation by an individual, as a result of his experiences in other situations. Sensitivity, as used in this study, is operationally defined as the ability to predict how another individual would describe himself on a number of selected personality traits. This definition is similar to that used in studies of empathy (7,2#,25,26,52,53) and social perception (3h,35, H3,MH,#§,67). The rationale for this definition is sug- gested by Luft (67): The judge...must take into consideration the possibility that the subject may cover up or exaggerate or falsify his "true" reSponses. In other words, he must be cognizant of the nature of the subject's ego-needs and his defenses, as well as the degree of insight which the subject possesses. But it is just these influences which the observer is at- tempting to discover in studying a subject. It therefore follows that the better the un- derstanding, the larger will be the number of items which the judge correctly predicts. ~16- Q-sort methodology will be employed in this study. The measure of a Judge's sensitivity will be termed his Accuracy score (ACC); the higher the ACC the more sensitivity pos- sessed by the Judge (J)—-i.e., the closer a J can predict the actual self-sort of a client (Other or "0"), the more sensitive will that Judge be deemed. Hypothesis 1 the sensitivity of an individual to nuances of a situation will be diminished if he ap~ proaches that situation with a preconceived categorization system In this investigation a Judge will be asked to do a predictive‘Q-sort1 based on, say, diagnostic material per- taining to an 0. He will then do a second predictive-sort based on, say, interview material of this same person. Hypothesis 1 states that the ”interview-sort" will be less accurate when it has been preceeded by the "diagnostic-sort“ than if it (the interview-sort) had been completed first. Rationale for Hypothesis 1 When a Judge does a predictive Q-sort he is performing a difficult--"complex and challenging"(13)--task. Other studies using this methodology find correlations between the predicted and the O's actual self-sort vary widely. For ex- ample: predictions based on hearing two successive tape re- 1 A predictive Q-sort involves sorting a p0pulation of Q- items, by J, as he thinks 0 would sort the same items. -17- corded therapy interviews (59) correlate from -.23 to +.37 with the actual self—sort done by 0; highly trained ther- apists attempting to predict the self-sorts of their own clients (26) have correlations from +.05 to +.85, with a median of +.H1; a Judge attempting to predict an 0's self- sort on the basis of diagnostic material (86) had correl- ations from .00 before therapy to +.55 at time of follow-up, 12 months after the therapy terminated. As a consequence of the difficulty of the prediction task, J, in an attempt to achieve some degree of success, must put himself in 0's "shoes", so to speak--an "imaginative transposing of oneself into the thinking, feeling, and act- ivity of another" (25). This process consists of, in part, forming hypotheses (66) based on the available material as to how 0 would respond to each Q-item. In this fashion the Judge builds up a category system or set concerning the Q-sorting behavior of O. The rationale, up to this point, is in accord with a more general statement of the Ames, Cantril, 23.§l trans- actional-probabilistic school of thought in perception: ...the organism has built up certain assump- tions about the world in which it lives.... All the presently existing assumptions of an individual taken together, constitute what may be called his "assumptive world". It is the only world he knows, and it determines the way he perceives the (physicalistic) world at any particular time. (2,p.278-9) -18- V Inkerms of the present study, a Judge builds up a sys- tem of hypotheses or assumptions about 0 based on the type of material presented to him. This "assumptive world" de- termines in large measure the predictive sorting of the J. After doing his first predictive Q-sort, the Judge is asked to do a second predictive sort for the same 0, based on dif- ferent material. He approaches this second task with a cat- egory system already formed. In other words, the Judge ar- rives at some "notions" as to what type of a person 0 is on the basis of the material presented initially. Thus, he ap— proaches the material on which he bases his second prediction with a preconceived categorization system. The above contention assumes that categorization sys- tems carry over from one clinical situation to another. Em— pirical support for this assumption is provided by Quinn (81), who found that therapists‘ attitudes are quite stable from client to client and are not affected by differences in the content of an hour or differences between clients (cf.7#,p.1). In accordance with Bruner‘s gating principle, in the situation cited above, the Judge in the second prediction session would be more open to cues or stimulus elements con- tained in the stimulus material which were in agreement with his preconceived notions; cues which did not fit his schema as well would be filtered out. Consequently, the Judge would not be making full use of all the cues available in the -19- second stimulus material. In effect, he is denying himself the use of negative evidence——negative in the sense that it runs counter to his preconceived categorization system. Therefore, the Judge's accuracy on this second prediction, as Hypothesis 1 suggests, should be less than if he had made this (second) predictive Q-sort without having completed one just prior to it.2 Experiments in perception (8,13,55,78,109) tend to sup- port the rationale outlined above. It has been found that once a stimulus has been categorized in a high probability, good-fit category, the threshold for recognizing cues con- trary to this categorization increases by "almost an order of magnitude" (11,p.131). Certain types of preconceived category stystems, how- ever, could result in increased sensitivity--namely, a cat- egorization system which was open to an "appropriate", rather than an inappropriate segment of the total range of stimuli in a particular situation. Such a system would be one which was properly attuned to the situation at hand: ...the most apprOpriate pattern of readiness at any given moment would be that one which would lead on the average to the most "veri- 2 There is a tacit assumption in this statement; namely, that the more "open to awareness" to all the cues in a sit- uation the Judge is, the more sensitive he will be; cf. Rogers‘ pr0position XVIII (84,p.520) and his discussion of therapeutic genuineness (87,p.97-8). -20.. BF“ :Lfiua "an“ _‘q 0 “h 'Zhrv I.“ “hvv‘ - cu: ‘L t‘,‘ fi‘ Vol r we... flhhl‘c-‘V‘V‘ A v» “9 V- not: dical" guess about the nature of the world around one at the moment.... And it follows from this that the most ready perceiver would then have the best chances of estimat— ing situations most adequately and planning accordingly. (11,p.130) Reasoning "backwards" for a moment, if an individual was able to demonstrate a high degree of sensitivity to a situation (in our terms, was able to predict the self—sort of O with a high degree of accuracy) then it could be said that he had an "appropriate" categorization system for that situation. Now, the general proposition states that the type of categorization system with which an individual ap- proaches an event influences his sensitivity to that event; consequently, if the statement is valid, the sensitive person, the accurate predictor, would have approached the situation with the "proper" pattern of categories. Hypothesis 2 is concerned with the nature of these "appropriate" patterns. Hypothesis 2 the sensitivity of one individual for another person will increase as the degree of con- gruelce between the categorization systems of the :wo individuals increases Rationale for Hypothesis 2 An individual's categorization system is indicative of the way in which he views the world (cf. pp.12—1h). If two persons tend to see the world in the same terms (with the same type of categorization system), then there can be said to exist an "empathic understanding" between these individuals. -21- _That is, since both persons perceive events in the same form, there will be an appreciation of the thoughts and feelings, each for the other. In short, the two individuals would’hee eye to eye' on things. If one of the persons described above is the O and the other is a Judge, then it could be expected that such a J would demonstrate a high degree of accuracy in his pre- dictiéns. Thus, an appropriate categorization system for a Judge in this study would be one that was similar to that of the 0. As Collins (50) writes: ...the formation of an impression of the personality of another is a large function of the underlying perceptual-cognitive organization process in the observer. It is highly unlikely that there can be perfect con- gruence between the categorization systems of two individuals; Hypothesis 2 holds that there is a direct relationship, how- ever, between the degree of congruence of the category sys- tems of the two individuals and the sensitivity of each for the other. Some experimental data supporting this contention had been reported (52,76). Notcutt & Silva (76) contend: The greatest success in predictability should be in those areas in which the greatest simil— arity of behavior or attitudes exists. The importance of the substantiation of Hypothesis 2 is suggested by Remmers (83) and Wolf & Murray (108): -22- The effectiveness of clinicians and counselors in general, can, as a matter of fact, itself be measured in terms of their ability to em- pathize. (83) One might hazard a statement that without empathy a man cannot make an accurate diag- nosis.... (108) Thus, both therapeutic and diagnostic competence is seen to rest on the apprOpriateness of the categorization system with which the clinician approaches these situations. To recapitulate: General Proposition: the type of preconceived categorization system with which one approaches events will affect sensitivity to nuances in those events Hypothesis 1: the sensitivity of an individual to nuances of a situation will be dimin- ished if he approaches that situation with a preconceived categorization system Hypothesis 2: the sensitivity of one individual for another person will increase as the degree of congruence between the cat- egorization systems of the two individ- uals increases -23- METHOD Overview In this study a group of clinical psychologists (Judges) was asked to predict the self-concept, as revealed by Q-sort methodology, of a client (Ophgg) who was being seen in therapy. The Input upon which these predictions were based, consisted of different types of stimulus material: (1) O's therapy interviews; (ii) diagnostic protocols ob- tained from O. In making his predictions the Judge fol- lowed Q-sort procedures. That is, J arrayed a pOpulation of statements, indicative of personality traits, along a con- tinuum of "most like" to "least like", as he believed the client would sort these same items. A Judge's score (Outake) was the degree of accuracy which he was able to approximate the client's actual sorting of the statements. The scores obtained by the Judges in this manner provided the dependent variable for this study. The independent variable was the sequence in which the stimulus material~~interview or diag- nostic—-was presented to the Judges. By presenting dif- ferent groups of Judges with various sequences, it was pos- sible to test (with analysis of variance) the effect of each type of stimulus material upon the accuracy scores obtained by the Judges. -24- Components of the EXperimental Design The italicized words in the preceeding section rep- resent the four major components of the experimental design employed in this investiagation (cf.26,#3,#5,97).1 The Judg --- whom the experimenter is attempting to measure The Other -—— whom the Judge is asked to investigate The Input --- information concerning 0 which is made available to the Judge The Outake —- the statements or predictions about 0 obtained from the Judge The Judges The individuals who served as Judges in this study were 32 clinical psychologists. They were employed either in the Counseling Center of a large midwestern university or by the Veterans Administration. These clinicians rep- resent a wide range of experience ranging from trainees with limited clinical eXperience to staff psychologists with 18 or more years of eXperience.2 1 In a further elaboration of this analysis, Gage & Cronbach (#5) classify investigations in this area along the following dimensions: (a) degree of acquaintance-~between J and O; (b) degree of extrapolation--amount of inference required between Input and Outake. In the present study the Js have little acquaintance with O and much extrapolation is required. In essence, the Judge must predict "hOW'gn individual deviates from the typical behavior of the particular group he belongs to--describing the unique behavior of the individual as in clinical diagnosis" (h5,p.%13). 2 See discussion in Controls, p.38. -25- The Other The individual, Mr. "L", who provided the stimulus material for this study was an undergraduate student at a large midwestern university. At the time that the study was undertaken he was being seen on a thrice-weekly basis by a therapist (not the author) at the university's Counsel- ing Center. Hr. "L", upon being asked by the author, agreed to participate in this research. He was told that he would be required to take several psychological tests (administered by the author) and that recordings of several of his therapy interviews would be needed. Mr. "L" was informed that his anonymity would be preserved, and, accordingly, all identi— fying material was removed from the test protocols and the therapy sessions. The Input The stimulus materials upon which the Judges based their predictions were of two main types: (i) diagnostic protocols, and (ii) interview material. The Diagnostic Protocglg consisted of the following: (a) the sex, age, marital status, academic class, and military status of the client (Hr. "L" was a Korean veteran) (b) a brief statement of Mr. "L's" academic interests (c) his percentile level on the Wechsler-Bellevue Intelligence Scale (d) a verbatim typescript of Hr. "L's" Rorschach performance; including, scoring, summary sheet, and location chart (e) a verbatim typescript of his performance on the Thematic Apperception Test; cards 1,2 1+,6BM,7BM,1O,12,13MF,13B,17Bl‘/l,18BI~vi, and 1?). (f) the comments and Mr. "L's" associations on the Draw-A—Person test. A copy of Mr. "L's" drawings were also included in the diagnostic material given to the Judges. (The client was asked to make his drawings on Ditto master sheets so that they could be reproduced.) ‘The Interview Material consisted of two recorded therapy sessions-~I' and I"--from which the therapist's remarks had been deleted. Verbatim typescripts, to be read while the recordings were played, were also provided for the Judges. The entire 11th and 13th therapy interviews were used. The therapist's remarks were removed from the recordings of the interview material for the following reasons: (a) this procedure would make the results obtained in this investigation more comparable to those reported in a similar study by Kell (59) (b) deleting the therapist’s remarks made the interview material more-"pure" in that the Judge could only hear what the client said and, consequently, would not be distracted by the therapist's voice, therapeutic acumen, therapeutic orientation, etc. (c) these deletions made listening to the inter— views a more novel experience, thereby in— creasing the Judges' attentiveness; further, it made the session somewhat more difficult to follow which had the effect of increasing the Judges' concentration on the material -27- (d) this procedure reduced the time required for listening to the recordings and, therefore, decreased the amount of time needed t9 coma plete the several tasks required of tne Js The method employed, and the considerations involved in selecting sessions I' and I" are described in Appendix G. By using both recordings and verbatim typescripts of the interviews he intent was to make the stimulus material as rich as possible. Use of the one without the other would have resulted in a loss of many cues available in the total situation. Support for the above contention is reported by Luft (68). This author finds a "distinct advantage" gained by listening to an interview as compared with reading a type- script of the same interview when Judges were asked to pre- dict O's reSponses on the basis of each type of material. Tupes (10%) also finds the use of both types of cue sources an advisable proceedural step. The Outake . In the sub-sections which follow the rationale for using Q-sort methodology and a description of how the Q- arrays were constructed will be given. Succeeding sub-sec- tions will contain a description of the scores obtained by this methodology and the interpretation given to these scores. h See discussion in Controls, p.h1. -28- A Q-sort Methodology was utilized in this study in view of the many advantages it was felt that it possessed. As Cronbach (18) has pointed out, this forced choice pro- cekdure is more "penetrating" than the common questionaire; t eliminates reaponse sets (6) such as saying "Yes" to all favorable items and "Ho" to all unfavorable ones; it is free from such idiosyncratic reSponses as saying "Cannot Say" twice as often as other reSpondants; it requires every person to put himself on the measuring scale in much the same manner; and, it frees the subject from many difficult and rather unimportant discriminations such as would be involved in ranking every statement. Block (9) compared a forced and unforced type of sorting and reports: ...the forced sort is at least as stable as the unforced sort.... The forcing procedure did not apparently decrease the reliable variance present. Using a simple distance measure, the forced sort emerges as more dependable.... More discriminations tend to be made in forced rather than unforced Q-sorting sit- uations. (9,p.#8h) Q-sort methodology also possesses a computational ad- vantage. Since Q-arrays have identical means and standard deviations (a consequence of the forced distribution), it is possible to compute Personian product-moment as by the quick, efficient method of differences procedure. 5 l 2 (16) r==1- D 2ng -29- Of the 32 Judges, only eight had previously partici- pated in research involving Q~sorting. Thus, this method- ology had the additional advantage of novelty which served to enhance the interest displayed by the Judges. The Q-arrays6 used in this study--Q' and Q"--each con- tained 70 statements. These items were derived from Edwards' (30) Personal Preference Schedule (PPS), which, in turn, was based on Murray's (75) system of needs. The decision to develop this particular item-population was based on the following considerations: Edwards has demonstrated (29,31) a .87 correlation be- tween the readiness to subscribe to an item as descriptive of one's self and the social desirability of that item. In view of this finding it was deemed inadvisable to ignore social desirability as a factor capable of influencing self- descriptive data. Consequently, an attempt was made to equate the two Q-arrays for social desirability.7 Since an equating on this variable had been achieved in the PPS, these itemswere selected. (Edwards' statements 6 See discussion in Controls, p.h1. 7 The writer attempted to obtain Edwards' original data in order to carry out this equating, but was unable to do so. In a personal communication Dr. Edwards referred the writer to a study by C. James Kleet (63). This work was obtained but, unfortunately, it did not contain the social desir- ability values for all the items. Consequently, the two Q- arrays were equated by a less precise method (see Appendix B). -30- were modified slightly-~usually by eliminating the words "like to"; e.g., "I(like to) participate in new fads and fashions"--in an effort to make these items more discrim- inable for the sorters.) A second consideration, which seemed to make these items particularly suitable, was the fact that they were based on Murray's personality theory. In view of the nature of this study—~particularly its diagnostic aSpects--it was advisable that items be used which were readily translatable into the types of hypotheses arrived at from diagnostic pro- tocols. Moreover, the inclusion in the test battery of the TAT--which was Specifically designed to elicit expression of these needs--seemed to make these particular items highly applicable.8 The two Q-sorts used in this study each contain 70 items. Sixty of these items were derived from the PPS. That is, each of Murray's 15 need categories is represented by four items in each Q-array. Ten additional items were added to each array since, in reviewing Stephenson's book on Q-methodology (95), it was noted that rarely was a state- ment population of less than 70 items utilized. Further, in selecting these additional items it would be possible to partially obviate a frequent criticism made of personality 8 It is possible that these items introduce a bias in the direction of aiding predictions based on the daignostic material; see discussion in Limitations, p.6h. -21- .J tests of the nature of the PPS. The question is frequently raised regarding the validity of the statements employed-- do they actually represent the needs to which they are assigned. with these considerations in mind, items were adOpted from Fordyce's dissertation (37). These statements had been selected for his study on an empirical basis9, as being rep- resentative of E (need) Independence and fi Dependence; they also had been sealed for social desirability. By incorpor- ating these items into the Q-arrays the social desirability balance was maintained and the sorts enlarged. In summary, then, each of the Q-sorts contained 70 items. In each array, for each of Murray's need categories, there were four items. Two additional need categories-~Independence and Dependence-~were represented by five items each, in both arrays. Thus, two Q-sorts, balanced for social desirability and possessing the advantageous "lumpy" quality (18)--clusters of items--were develOped; Appendix D lists the items of both Q-arrays. The Scores10 obtained using Q-methodology depend on the 9 Fordyce selected his statements in the following manner: 307 items were submitted to seven clinical psychologists who classified the statements as to "independency" and "dependency" (the clinicians supplied their own definitions of these terms). Two hundred and twenty-five (225) of the items were classified identically by five of the seven Judges. These items were then rated for social desirability by 68 college students. 10 The data of this study consist of correlation coefficients which were transformed into "scores" by using Fisher's‘g' transformation; cf.28,p.126-8). -32- number and kinds of arrays the Judges are required to make. In this investigation each Judge did three sortings: i. a self-sort--in which the J distributed the items along a continuum of ”least like" to "most like" for himself ii. a predictive-sort--in which J placed the statements along a'least like” to "most like" continuum as he believed 0 would have sorted these same items iii. a second‘predictive-sort--as in (ii) with J using the alternative from Q-array11 From these sortings several scores were attainable: a. Accuracy (ACC)--also called "prediction score" (3%) and "raw empathy" (7,53)--this is the correlation between J's predictive sort and 0's actual sort b. Real Similarity (RS)-—cf.3#--the correlation be- tween J's self-sort and 0's self-sort c. Attributive Similarit (AS)-—also known as "pro- jection" 7,53); "assumed similarity" (#2); and,_ "non-differentiation" (H)--correlation between J‘s self-sort and his predictive-sort For example, Judge A (see p.37)‘would have the following scores; the notation ":” means "correlated with". a. ACC--(1) predictive Q' : 0's Q' (l) predictive Q" : Q's Q" b. RS---J's self-sort (Q") : 0's Q" J's predictive-sort (Q")12 0. c. AS--~J's self-sort (Q") 11 Each Judge used both T~arrays in one of two sequences: (1) Q" ', Q"; or, (ii) Q', Q", Q'; of. discussion in Controls, p.h1. 12 A second AS score could be obtained by correlating J's self-sort (Q") with J's predictive-sort (Q') using the 17 "needs" as the data to be correlated; of. discussion in Results and DiscussiOn, p.#5. -33- The Interpretation of the Scores was as follows: ACC is essentially "understanding the other person in the diagnostic sense of the word" (3H,36). Gage, gt a; (#6) list the following terms for ACC: diagnostic ability, empathy, interpersonal perception, sensitivity, social per- ception, and understanding. These terms have a common ele- ment in that they convey the idea of "clinical intuition" to use Hathaway's expression (5%). In this study the term "accuracy" will be used as indicative of sensitivity. That is, the higher the ACC the greater the sensitivity displayed by the Judge. BS is defined by Fiedler (3k) in the following manner: "Real" itlthiscase indicates, of course, only the real similarity between self-descriptions, and only to the extent to which the self des- criptions reflect personality structure can we speam of real similarity between the per- sons involved. AS has been used frequently in studies of this type, but, as the plethora of other names may suggest, it is the most difficult of the three scores to interpret. A more extensive discussion of this score and the reasons for its exclusion from this study are given in Appendix H. Procedures Since the procedures for O and J differ, each will be Presented in a separate sub-section; a summary of these Procedures, in tabular form, will conclude this section. -34. Procedure for Other The recorded interviews were obtained in the manner des- cribed in Appendix G. Tn the testing session 0 was given a sheet containing suggested instructions for doing a Q-sort; Appendix C. He did a-self—sort using the items in Q'; he then did a second self-sort using the items in the Q" array. The required distribution for these sortings is shown in Table I . TABLE I REQUIRED FREQUENCY DISTRIBUTION FOR ALL Q—SORTS Pile Number 1 2 3 H 5 6 7 8 9 LEAST LIKE MOST LIKE Number of Items 1 3 7 15 18 1S 7 3 1 Following the second self—sort 0 was administered the projective tests (Rorschach, TAT, DAP). 0's final task was to rate himself on each of the 17 need categories on a nine point Rating Scale 13; Appendix E. Procedure for Judges Each Judge was asked to complete an Experience Data Sheet (Appendix F). On the basis of the information obtained 13 This scale was composed of the definition and list of defining adjectives used by Murray (75) to characterize each of his need categories. -35- he was assigned11+ to one of the experimental conditions; this determined the sequence in which the stimulus material-~diag- nostic or interview--was presented to the J (see below). After completing the Experience Data Sheet, each Judge was given the suggested instructions for doing Q-sorts (Ap- pendix C). The J then did a self-sort. Following the pre- sentation of the first stimulus material, the Judge did his first predictive-sort. Then the second stimulus was pre- sented and J did a second predictive-sort. Following this third sort, the Judge was given the Rating Scale (Appendix F) and was asked to make a "diagnostic" rating of Mr. "L". That is, the Judges were instructed: "Now, on this sheet, rate the subject as you think he really is." To summarize, the procedures employed were: Other 1. self-sort with Q' items ii. self-sort with Q" items iii. projective tests iv. self-rating on Rating Scale i. complete Experience Data Sheet ii. self-sort* . iii. first stimulus material presented iv. first predictive-sort* v. second stimulus material presented vi. second predictive-sort* vii. diagnostic rating of O on Rating Scale * Sequence (1) or (ii), page 33, was followed 1% 'Originally the intention was to match Js on all items listed on the EXperience Data Sheet. However, this procedure was infeasible due to the limited number of clinicians avail- able to the writer. Matching was based on only therapeutic and diagnostic experience; see discussion in Controls,p.38. ~36- _w Experimental Conditions The four experimental conditions employed consisted of different sequences for the presentation of the stimulus material. These sequences were: i. Interview material followed by Diagnostic (I"D.) protocols ii. Diagnostic protocols followed by InterView (D I") material iii. the 11th Interview followed by the 13th (1'1") Interview iv. the 13th Interview twice (I"I") Figure 1 presents these sequences in schematic form. In addition to the symbols already introduced,l stands for "pre- dictive Q-sort based on Interview material"; 2 means "pre- dictive Q-sort based on Diagnostic protocols"; capital let- ters represent the Judges. Session "two" immediately fol- lowed session "one". ONE Sessions TWO E I II I II I x C i. A--D E-—H I" ' Q A-—D E--H p O ————————————— ' —————————————— e n I. II I II I r d ii. I--L h--P 2 ‘ I" I-~L h-—P i l _____________ ' ______________ m ( I u I n t e 1 iii. R--U V—-Y I' ‘ I" R—-U V--Y n o _____________ ' ______________ t n I II I II I a s iv.AA--DD EE——HH I" ' I" AA—-DD EE--HH l l Figure 1.‘ Schematic presentation of procedures for all Judges; N = 32, n = 8. -37- Experimental conditions I"D and D I" (i and ii in the above schema) indicate that the interview and diagnostic material were presented to the Judges in a counter-balanced sequence. This design permitted the evaluating of the effects of one type of stimulus material on the other. Experimental condition iii--I'I"--presented the Judges with the same kind of material (all interview). This was an attempt to control for the "heterogeneity" of the stimulus material in conditions i and ii, while keeping the'ambunt~ of information'given the Judges roughly the same. Experimental condition ig--I"I"--presented the Judges with the same material on two occasions. This was an attempt to control for the fact that the Judges in the other conditions were given different material to base their predictions on, while keeping the "amount of exposure" to the stimulus mat- erial roughly the same. Additiongl700ntrol Features In addition to the controls incorporated in the experi- mental conditions, an attempt was made to control other factors in an effort to increase the rigor of the investi- gation. These factors, their rationale, and the procedures utilized will be detailed in this section. (a) Experience-~Kell (59), in a study somewhat similar to this one, used experienced therapists, pre-practicum students, and chemists (n=16). He found that there was a -38— “a significant difference between the mean accuracy of the predictions made by these groups. In "step-wise“ fashion, the therapists were significantly more accurate than the students, who, in turn, were significantly better than the chemists. In view of this finding, and the wide range in ther- apeutic (and diagnostic) experience present in the popu- lation employed in this study, it was deemed necessary to control the experience factor. Since diagnostic protocols as well as therapy interviews served as stimulus material in this study, an attempt was made to take cognizance of both types of clinical experience.15 TABLE II RANGE OF JUDGES' CLINICAL EXPERIENCE UNDER FOUR EXPERIMENTAL CONDITIONS Congggion* n Therapy** Diagnosis*** I"D 8 37 — 5000 12 ~é- 1100 D I" 8 0 - 5000 6 ~-- 2000 1'1" 8 o e #500 30 --- 800 I"I" 8 35 — 5000 0 --- 1000 * See Experimental Conditions,p.37ff. ** No. Individual hours conducted and supervised *** No. Diagnostic reports completed and supervised 15 Based on information Js provided on Experience Data Sheet, each was assigned to one of the experimental conditions; care being taken to have same number of highly experienced peeple in each group. -39.. Since the data in Table II is of an approximate nature, in order to ascertain whether the groups of Judges differed significantly on the therapy— and diagnostic-experience vari- ables, a non-parametric test was employed--Kruskalewallis H- test (analysis of variance using ranked data) (91). The results of this analysis16 led to the conclusion tiat the four groups did not differ significantly; i.e., the groups of Judges were reasonably well equated in terms of amount of therapy- and diagnostic-experience. (b) Time Spent with Stimulus Materialm-Since two kinds of stimulus material were used—~diagnostic and interview-~in order that differences in ACC which might occur could be properly attributed to the stimulus difference, it was deemed necessary to control this "exposure” factor. That is, if the Judges in one experimental condition spent less time with, say, the interview material, then it would be unclear whether differences which might ocCur between the ACC scores of this group and the scores of another group, were due to the ex- posure factor or the experimental condition. In accord with the above considerations, a minimum ex- posure time was set by the writer. The playing of the deleted-therapy~interviews (cf.p.27ff.) required approxi- 16 The obtained values for the therapy- and diagnostic— experience variables were -.956 and -.3.0397, reSpectively. For three degrees of freedom a value of 7.815 is required for significance at the .05 level (11.3%1 at .01). -ho- ..I mately 20 minutes. A Judge was not permitted to begin sorting until after the interview had been played thrOugh. This same minimun time—-20 minutes-ewes used with the diagnostic material. Each J was told he might spend 20 minutes, even longer if he wished, before beginning his predictive-sort. It was possible to maintain this minimum exposure time by with- holding the pile of Q-statements until this time had elapsed. (c) Fatigue-~The modal time needed for the Judges to complete all the required procedures was two hours (the range was from 85 minutes to three hours and 20 minutes). While little could be done to control these individual differences, it was felt that procedural steps which would help decrease this amount of time might keep fatigue at a minimal level. The decrease in the amount of time required for hearing the therapy sessions engendered by deleting the therapist’s remarks from the recordings, was one method used to control fatigue. It was felt, also, that these deletions increased the novelty of the procedures and that this too would serve to keep fatigue at a lower level. Another innovation, which also had the goal of helping to reduce fatigue, was the use of two equated, though different Q-arrays. It was hOped that alternating the packs of Q— statements would decrease boredom and, thereby, fatigue. (d) Memory~~The use of two Q—arrays, an experimental device unique to this investigation, also served a function -h1- other than that mentioned above. Previous studies are re- ported in the literature in which the Judges or Others were required to do more than one Q-sort-~e.g., the self-, ideal self~, and ordinary person- sorts that are frequently used. However, in these investigations the same pOpulation of statements is utilized for all sortings. This procedure introduces an uncontrolled, possibly differentially Opera- tive, confounding factor; namely, memory. In such an eXperimental design, a Judge, possibly re- membering some of the item placements of his first sorting, might consciously attempt to array these same items either the same or differently in subsequent sorts. It becomes unclear whether the correlations which would be obtained under these circumstances are due to the experimental con- ditions employed, or are a function of the Judge's memory --be it accurate or not-~or some combination of the two. The literature does not contain any introspective data obtained from Judges on this issue. In introducing a second, equivalent, population of Q- statements, the intent is to: 1) minimize, to some degree, a confounding memory factor 2) maintain interest at a higher level by deminishing the effects of boredom 3) permit the obtaining of a reliability measure for O -1}. 2... ‘It was these considerations which led to the decision to deveIOp two Q-arrays.' The two piles of statements were given to the Judges in a mixed sequence--(i) Q", Q', Q"; or,(ii) Q', Q", Q'--with the expectation that this procedure might interrupt any memorization process. In an effort to further minimize the differential in- fluence of memory, a second control was employed. The Js were permitted to take notes during the time they were ex- posed to each type of stimulus. They retained these notes, as well as the printed stimulus material, while doing their predictive Q-sorts. In other words, a Judge was free to "refresh" his memory of the stimulus material while making his predictions. These same conditions prevailed for both prediction ses- sions; however, following session one, all notes and printed materials were removed prior to the presentation of the stimulus material for prediction session two. (e) Interaction between Q-sort§--Using two Q-arrays emphasizes the possibility that the items in the first array may interact with those in the second one. This possibility is also present when the same pOpulation of-Q—statements is used for all the sortings, but it appears to have been rel- atively ignored as a source of "error" in previous researches using this methodology. Through the use of a counter—balancing technique, an -h3- attempt was made to "cancel out" interaction effects. That is, four Judges in each experimental condition used sequence 3 (see section "d" above) and the other four Judges used sequence ii. Since the main concern of this investigation was with group differences produced by the different experi- mental conditions, this procedure served to control possible interaction effects by permitting such effects to Operate to the same degree in each of the experimental conditions. (f) Social Desirability of Items in Trait POpulation-— See discussion on page 30ff. and Appendix B. RESULTS AND DISCUSSION Cronbach (18) points out that data obtained with Q- arrays in which there are "clusters" of items are suscept- ible to two different types of analysis; correlations may be computed on the individual items, or the "scores" for each cluster can serve as the basis for the correlations. The latter method gives more reliable measures and is preferred. In the present study the clusters of statements represent need categories. The analysis was conducted on the correla- tmn of these "need scores".1 That is, the mean of the item placements for a particular need category was taken as the score for that need.2 Thus, 0 and each Judge had 17 scores for each Q-sort. Table III (p.h9) shows the results of the computations carried out on these cluster scores. Reliability Other A reliability coefficient for O was computed using an analysis of variance technique. Such a procedure makes it 1 Correlations were also computed on an item basis. The results of this statistical treatment are given in Appendix J. 2 This procedure, although routinely employed in studies of this nature, makes a tacit assumption regarding the clusters of items; namely, that each cluster comprises a sub-test in the total pOpulation of items and, further, that the integrity of these sub-tests is not destroyed When the cards (on which the items are printed) are shuffled prior to being sorted. -45- possible to ascertain whether or not the test measures with sufficient accuracy to differentiate between the needs; and, also whether or not the practice effect is significant. Table IIIshows the results of these computations. TABLE III RELIABILITY'FORT"0THEB” I sonnet: or ' SUM or ‘ 1mm VARIANCE SQUARES g; SQUARE g 2 Practice Effects ' (Between Q’ a Q") .0002 1 .0002. ~— Between Needs 17.#6’+’+ 16 1.0915 2.6361} .05?p_>.91 Error. 6.6221 _1__6_ .Mho ' roux. ~ . 2150897 33 g ' 2 Between reeds - Error =- .l+5o ' ' ‘ tt Between fieéas + Error J -’ *. Admittedly, the reliability coefficient obtained for O is low; however, the analysis has demonstrated that the Q-— sorts, as used by the 0, differentiate between the various need categories at better than .05 level. Regarding the advisability of using a criterion with low reliability, Thorndike writes: ...it is not of critical importance that the reliability of a.crIterIon be high as long as it is.esteblished as definitely greater than -ng- zero. Even when the reliability of a cri- terion is quite low, given that it is def- initely greater than zero, it is still pos- sible to obtain fairly substantial correl- ations between that criterion and reliable tests and to carry out useful statistical analyses in connection.with the prediction of that criterion. (101,p.106) Since an‘g = .hSO does not reach the value (.H68) re- quired to be significantly different from zero at .05 level (NA: 17), the results of this study become more equivocal.3 Judges The consistency with which the Judges used the two Q- arrays was computed following a technique outlined by Ebel (27). The coefficients obtained-~73? and .796 for Q' and Q", respectively--are each based on the first predictive sorts of eight Judges; all 16 Judges made their predictions on the basis of l". The 16 J5 used in computing these reliability coef- ficients were the eight assigned to the I”D condition and the eight assigned to the I"I" condition. The first four Js in each of these groups made their initial predictions using Q'; these sorts were used to compute the reliability of Q‘. The second four Judges in each of these groups made their first predictive sorts using Q"; accordingly, the reli- ability for Q" was based on these arrays. 3 A reliability coefficient for 0 based on the individual items was also computed; r = .197. Thus, the advisability of using the "need scores" and the validity of Cronbach's recommendation (o.h5) was demonstrated. -h7- "H The reliabilities obtained in this study are lower than those found by Frank (39). This author, using a test-retest methodology, reports correlations ranging between .93 and .97 for self-sort with Q—statements. The lower reliabilities found in this investigation may be the result of (i) basing reliabilities on predictive-sorts rather than self-sorts; (ii) using two trait-pOpulations rather than the same one for all sortings.h Results In order to properly evaluate the data obtained in this investigation it is necessary, first, to ascertain whether or not the Judges were able to predict 0's self-sort with a degree of accuracy greater than chance. If the Judges were unable to exceed a chance level, then any further statistical analysis would be open to serious questioning. The correlations reported in Table IV (p.h9) reveal that of the 6% ACC scores, 20 were significantly different from zero at the .05 level (12 of these at .01). A further analysis of this data, shown in Table V (p.50), indicates that 16 Judges had at least one ACC score signifi- cantly different from zero at the .05 level; 11 Judges were able to reach the .01 level on at least one prediction. # See discussion in Controls, p.#1ff. ~48— FOR ALL JUDGES TABLE IV” ACC AND RS CORRELATIONS '8l':" 8'8‘::"'SESSION GP. GROUP JUDGE ONE ADC TWO MEAN RB ' A .032' -.113 .563** B .717** -.029 .530* C‘ .412 .H47 .502* D .5.7* .352 .393. I"D E -_. -..0 58 .523* F .783 ** o “‘6 5 ~51.- 5 I”? G .601** .331. .375 H .108 .0 ' .9208 Mean -- .4hh** --- .11 .312** I -.27% .080 .472* J .082. .I21 .557* K e 311'? “0 021+ .o 272 L .593** .388 .643“ D I“ M .759** .10#’ .295 N .gg8** .gg5* .H06 0- . 9 . :6: .691** P! .6#5** .M75* .733** Mean "" .)+50** "'"" o 279** 0367** R. .e 639*," o 299 .0 533* S .567* .536* .735** T .581* .17% .279 U .09H .158 .750** 1‘1" V -.217 .h33 .183 W o 667** 0376 -0005 X .733** .628** .380 Y .E+B* “0033 0031+. iMean -- . 82** -—- .3 7** .Ml6** AA .6#l** -.0 1 .256 BB .#07 .#15 1609** CC .079 .#29 .11% DD .105 ~.0#5 .135’ I"I" EE .559* .205 .669** FF -.1065 .429 .655** GG .298 .368 .293“ HE .391 .122 .036 Mean -- .316** --- .239** .278** .k27** .262** .3n7** TOTAL MEANS *“Signiricantly different ** Significantly different from zero at .05 level from zero at .01 level -59- TABLE V NUMBER OF JUDGES ACHIEVING ACC SCORES DIFFERENT FROM ZERO AT .05 AND .01 I LEVELS OF SIGNIFICANCE Level of Number of ACC Scores Significance at this Level 0 1 2 .05 16 12 1+ .01 21 10 1 It is difficult to evaluate the frequencies reported in Table V since there is no way of determining how many ACC scores, different from zero at a specified confidence level, could be expected from the group of Judges by chance alone. Moreover, the question arises as to the magnitude of ACC scores which could be expected to occur by chance. In an attempt to establish a base level from which the obtained ACC scores might be evaluated, a "robot" was constructed. :Bobot" Accuracy A self-sort and two predictive sorts, using the two Q— arrays in sequence i—-Q", Q', Q"--were made for the Robot. That is, for each sort the pile of cards containing the 70 Q-statements was shuffled and arrayed into the required frequency distribution (cf.p.35) according to the dictates of a table of random numbers. The obtained scores were: ACC (Q') .017 RS -.012 A00 (Q") -.103 -50... _ Following the procedures outlined by Edwards (28,pp.128— 131) fiducial limits at the .05 level were established for the Robot's scores. These were: ACC (Q') +.356 to -.338 ACC (Q") +.276 to -.#32 RS (Q") +.3h2 to -.360 If the ACC scores in Table IV were evaluated with the fiducial limits of the Robot's scores in mind, the results are as shown in Table VI. TABLE VI NUEBER OF JUDGES ACHIEVING ACC SCORES BEYOND ROBOT'S FIDUCIAL LIMITS* Number of Times Number offi Robot's Scores Judges Exceeded o 8 1 11 2 13 * Fiducial Limits established at .05 level The results shown in Table VI demonstrate that 75% (2% out of 32) of the clinicians, who served as Judges for this study, were able to predict the self-sort of a client, at least once in two trials, significantly better than chance; slightly better than h0% of the Js exceeded chance on both predictive sorts. -51- In addition, there was a highly significant degree of agreement among the Judges when they rated 0 "dia'nostically" (cf.p.35ff.). These ratings were ranked and Kendall's (91) coefficient of concordance computed. The value obtained, 1%7.2h, for 16 degrees of freedom, is significant beyond the .001 level. A corrected rank-difference correlation of .H19 was found between C's own ranked ratings on the Rating Scale and the consensus ranking of the 32 Judges. This value does not reach the level required for a significant difference between a rank correlation and zero (102). The magnitude of this correlation, however, suggests that the role played by diag- nostic ability in a Judge's ACC score would be a fruitful area for investigation. The above results were interpreted as indicating that the Judges in this study predicted better than the level which could be expected by chance alone. Further, there was a significant degree of agreement among the Judges in the impression they gained of 0 from the divers stimulus material. Hypothesis 1 Hypothesis 1 holds that an individual's sensitivity to a situation is diminished when he approaches that situation with a preconceived categorization system. In attempting to demonstrate the validity of this statement it is necessary \ -52.. -49. i_-_~_.._._i—.i_ A to take into account three variables involved in the study: i. Experimental Session -— 'one'and 'two' ii. Type of Katerial -- diagnostic and 'nterview iii. Sequence of haterial -- heterogeneous (I"D 'D I") ' homogeneous (I'I" identical (I"I") Support for Hypothesis 1 would obtain if the ACC scores in session 'two‘ were significantly lower than those in ses- sion 'one'. Since it was possible for interactions to occur between the three variables, analysis of variance was selected as the apprOpriate statistical technique. Through this type of analysis it was possible to test all the variables and their interactions, simultaneously. Following the method outlined by Edwards (28,pp.288-296) for analysis of variance with repeated measurements on several independent groups, the results shown in Table VII (p.5h) were obtained. The E between sessions of 7.35##, for 1 and 28 degrees of freedom, is significant between the .05 and .01 levels. An examination of Table IV (p.h9) reveals the total mean for session 'one' to be higher than the mean of session 'two'. Consequently, it was concluded that the categorization system which the Judges brought into the second prediction session led to significantly lower ACC scores.5 To the extent that 5 The ACC scores of 23 of the Js in this study showed this trend. Of the nine Judges whose predictions were in the op— posite direction (i.e., the second prediction more accurate than the first) not one td‘their ACC scores was significantly different from zero. -53- these ACC scoretreflect the Judges' sensitivity, Hypothesis 1 has been demonstrated. TABLE VII ANALYSIS OF V'm RIANCE OF ACC SCOBEQ FOR FOUR GROUPS OF JUDGE S UNDU‘ DIFFER“ SNT EXPERILENTAL CONDITIONS WIl mH TWO SEODIOTN EOR EACH GROUP SUM OF LEAN SOURCE 'i ViL'tIAIION S‘UARES g: S’UARE E Between Conditions .2209 3 .0736 -—- Between Js in same Condition 3.6083 g8 .1288 Total Between Js 3.8292 31 Between Sessions .5685 . 1 .5685 7,35hh Interaction: Ses- sions x Conditions .0831 3 .0277 -_- Interaction: Pooled Js x Sessions 2.16 2 28 0773 Total Within Js 2.8868 32 TOT ”AL 6.6%60 63 The finding of a significant difference between sessions, although in accord with the gating principle in Bruner's (11) general formulation, is also embarrassing to the part of his theory dealing wi h the "closer look". Bruner contends that permitting an individual to take a "closer look” at a situa- tion is one of the antidotes for non-veridical perception. Experimental Condition I"I", in essence, provided eight of -5u- the Judges with this "closer look". As Table IV (p.h9) reveals, the group mean for I"I" is higher for session ‘one‘ than it is for session 'two'. While this difference in not statistically significant (3 = 1.3735; g; = 7; .50>p>.20) it is in the Opposite direction to Bruner's theorizing. Ac- cording to the ”closer look"principle, the Judges should have higher ACC scores in session 'two'. The data of this study appears to be more in agreement with Hathaway's statement: It would seem that the consistant error of both individual participants and of groups somehow derives from having too much infor- mation—~at least too much information de- sturbs the predictive process. ...Of course the arguement would be absurd if the ad- ditional information were directly informa- tive about the items to be predicted. (5%) In an attempt to isolate the experimental conditions con- tributing to the obtained significant difference between the two predictive sessions, separate analyses of the individual experimental conditions were conducted. At the same time, the possible effect of different levels of therapeutic and diagnostic experience was investigated; these tables are given in Appendix I. Parenthetically, it may be noted that the significant relationship between level of experience and predictive ability reported by Kell (59) was not replicated in the results of this analysis. The tables in Appendix I reveal that no one experimental -55- condition resulted in a statistically significant difference between session 'one' and session 'two'. However, the Es obtained for the I"D and D I" conditions are consistently higher than those found in the 1'1" and I"I" conditions. In an attempt to ascertain whether the heterogeneous or the homogeneous material differentially affected the Judges' performance in the two sessions, two statistical analyses were conducted--Tables VIII and IX. TABLE VIII ANALYSIS OF VARIANCE OF ACC SCORES APPLIED TO I'I” AND I"I” CONDITIONS SUM OF MEAN SOURCE OF VARIATION SQUARES g: SQUARE 3 Between Conditions (I'I" & I"I") .1985 1 .1985 2.h320 Between Js in same Condition 1.14h6 lfl’ .0817 Total Between Js 1.3k31 15 Between Sessions .1hh8 1 .1h#8 1.5210 Interaction: Ses- sions x Conditions .020# 1 .020# --— Interaction: Pooled Js x Sessions ‘1. 28 13 .0952 Total Wi thin Js 1 .4280 lg TOTAL 2.8%11 31 ~56— Neither of the Es obtained in Table VIII, for 1 and 14 degrees of freedom, reach the value (4.60) required for sig- nificance at the .05 level. Thus it would appear (and Table IX bears this out) that the lowered predictive accuracy in session 'two' is a con- sequence of the interference produced by the presentation of heterogeneous stimulus material. In order to investigate more closely the divers effects of the remaining two vari- ables (p.53) on the heterogeneous conditions, a statistical methodology for counter-balanced tests (9%) was employed. Table IX shows the results of this analysis. T2113 LE I} ABALYSIS or 'ARIANCE FOB COUNTER-BA LNOED TESTS APPLIED TO I"D AND D I" CONDITIONS w —u—— SUM OF MEAF SOURCE OF VARIATION SQUARES g; SQUARE pg Independent Results: Presentation order .0221 1 .0221 ~-- Residual Between Js 2.M6EZ .fli. .1759 Total Between Js 2.h858 15 Correlated Results: ' Between l" and D .0160 1’ .0160 --— Between Sessions .h703 1 .h703 7.9175 Residual Within Js .8325 1&_ .059H Total Within Js 1.3188 lé TOTAL 3.80%6 31 -57- As shown in Table IX, accuracy of prediction is not significantly affected by the sequence in which the material is presented,6i.e., whether I" preceeds or follows D; nor is it affected by the type of material upon which the prediction is based--I" or D. The E for Between Sessions of 7.9175, for 1 and 1% degrees of freedom, is significantly different from zero at the .05 level. Since this result is based on the effects of making two predictions on each Judge, it becomes apparent that it is something within the J which accounts for the obtained result. To the extent thatthis interference reflects the operation of a gating process, it may be con- cluded that the phenomenon is most pronounced when there is a transition from interview to diagnostic material (or zigg XEEEE>° Thus, this statistical procedure lends added weight to the contention that the differences found between pre- diction sessions are the consequences of a gating process. 6 An alternative statistical analysis, suggested by Dunnett's (23) procedure for comparing several treatments with a standard, provided additional support for this conclusion. The mean A00 for each session in each of the experimental conditions was computed. In three of the conditions the ACC based on I" is preceeded by D, I', and I"; in the re- maining condition (I"D) the ACC based on I” was the first prediction made. Thus it becomes possible to compare the effects of three different types of preceeding material on I" with a standard—-I” as the first prediction. The largest difference between a treatment mean and the standard was .2335 (for I"I"). Since a difference of 3.1390 is required for a one-tailed test of significance at the .05 level, it was concluded tha none of the three sequences had a differentially significant effect on ACC. -58- Hypothesis_§ Hypothesis 2 states, in essence, that if two people’See" themselves in a similar way, then the amount of sensitivity of these individuals, one for the other, will be greater than if this initial similarity did not exist. RS was in- terpreted as indicative of the amount of similarity existing between 0 and J (cf.p.3h). Thus, the confirmation of Hypoth- esis 2 would require that a significant correlation be found between ACC and RS; i.e., the more sensitive the J tie higher his RS score. Since the analysis of the ACC data revealed that there was a significant diff rence between the scores in the two prediction sessions, a mean ACC score was computed for each Judge; these means were then ranked. A rho correlation be- tween these ranks and the ranks based on the Judges' RS scores was calculated. The obtained value, 3 = .201,did not reach the required value for a rank correlation (N=32) to be significantly different from zero (102). In an attempt to ascertain whether or not R8 was sig- nificantly related to the extreme ACC scores, a biserial cor- relation was computed. On the basis of their mean ACC scores the Judges werc separated into two groups-~the upper and lower 27% (n29). A biserial r was computed from the RS scores of these 18 Judges. The obtained value, 3 = .207, was not sig- nificantly different from zero. -59.. Although both of mlese results are in tide predicted direction, in View of the small magnitude of the obtained correlations, it must be concluded that RS is not a sig- nificant factor contributing to high ACC scores. Thus, the results of this analysis do not lend statistically signifi- cant evidence for the validity of Hypothesis 2.7 Viv—1v.— To recapitulate, the statistical analysis employed in this study has led to the following conclusions: Diagnostic material does not markedly enhance nor signi.ficantly decrease the sensitivity of a Judge to Interview material; and, similarly, for the effect of Interview material on Diag- nostic material ‘ Although the presentation sequence of the stimulus material does not significantly ef- fect sensitivity, making two predictions, partscrlSrlv if based on wid. ely differing types of 1%. terial, results in significantly less accuracy for the second prediction O1. t was found for the validity of Hypothesis 1 h validity of Hypothesis 2 rema ns in question; h results were in the anticipated direction but oi- not reach the level of statisti al signifi- cance ‘JC‘I‘i—B CJQGD \ The general prOposition underlying this study, was, in the main, supported by the above find rings ‘ .. 7 The experiment3l procedures employed in this study may not have tested this hypothesis adequately, since the pos- ible influence of iicat was not controlled. That is, it ‘may may have been the ca se that some of the obtained inac— curacy in ACC is attributable to a disinclination on the part of the Judges to "see” (predict) O as simile r to them- selves; cf. discussion p.68. -60.. Limitations As stated previously, a broad purpose of this investi— gation was to examine certain variables that might be re- lated to a therapist perceiving his client more (or less) accurately, and which, in turn, would influence the effective- ness of the therapeutic relationship. We have not, however, studied psychotherapy directly. Consequently, the results of this investigation are applicable to psychotherapy only to the extent that certain assumptions are accepted. A first assumption is that an awareness on the part of the therapist of his client's phenomenological frame of ref- erence is an important variable in effective therapy. Such an assumptbn is directly in accord with the tenets of Client-Centered therapy, as outlined by Rogers (8%,85), in which great emphasis is placed on the necessity of an empathic understanding existing between therapist and client. - From a more analytical orientation, Frieda Fromm-Reich- mann writes: We know that the success or failure of psychoanalytic psychotherapy is...greatly dependent upon the question of whether or not there is an empathic quality between the psychiatrist and the patient. (N1,p$62) Otto Fenichel, writing from a classical psychoanalytic viewpoint, cites as a first rule regarding interpretation: One should always start the interpretation at the surface" (32,p.hh). He goes on to say: M— We interpret, as is well known, what is al— ready in the preconscious--and just a little bit more—-which thereby becomes capable o entering consciousness. (32,p.53) Whether it serves as a basis for "empathic understanding" or for determining what constitutes "just a little bit more", the foregoing suggests that the phenomenological world of the client is regarded as an important variable by major theori— ticians writing on psychotherapy. A second assumption, concerning Q-methodology, is re- quired before the results of this study can be interpreted as having meaning regarding effective psychotherapy. It was assumed in this research that Q-sorting afforded a means for obtaining a quantifiable, flexible (18), highly complex and multi-dimensional (6%) View of a client's phenomenOlogical world. Several studies in this area (38,6%,87) find that Q-Sorts can adequately reporduce the phenomenology of the sorter; and, further, that a Q-sort does seem to be able to say what the sorter wants it to say despite his doubts that his impres— sions are "coming through”. To the extent that one is willing to grant the validity of the above assumptions, the results cited in the preceeding section have relevance regarding the usefulness of pre-ther- ‘apy formulations (based on diagnostic tests)in enhancing therapeutic effectiveness. Nevertheless, these results must be interpreted in light of the limitations imposed by the ex— perimental design employed in this investigation. -62- In this study many Judges attempted to predict the self~ concept of one 0. The procedure restricts the degree to which these results may be generalized. Questions might be raised regarding the "typicality" of O; i.e., was the 0 used in this study typical ofthe clients seen in therapy? Would the same results obtain if a female client had been used? If female Judges had been used? What if a more disturbed, or a more reliable client had been selected; would the same con- clusions have been reached? Suppose the Judges were required to predict the self-sorts of two very "different" Os; would this have changed the results? The present study provides no answers forthese questions. The moderate reliability of the Q-arrays used in this investigation imposes another limitation on the obtained re- sults. One wonderswhether the magnitude of the ACC scores might not have been higher with more reliable instruments. If so, could this have changed some of the findings? Might not some of the obtained insignificant Es reach the level of statistical ignificance with a more reliable measuring in- U) strument? In other words, the relative unreliability of the Q-sorts may have produced Type II errors (cf.28,p.29-30). A further limitation in this study involves the Judges. It has been pointed out (#6) that correcting scores through the use of "intermediary keys"--e.g., J's stereotype of the "typical" client—~can make studies of this nature more rig- ~63- orous. Cronbach & Gleser (19) suggest statistical proceédures which can be used when Js are asked to predict the self-sort of a "typical client", along with their other predictions. This procedure was not employed in this study. It was deemed inadvisable to require the Judges to make still another Q-sort; the time factor, which tended to increase the likeli- hood of disruptive effects due to fatigue, was becoming too formidable. horeover, the issues raised by Baker & Sarbin (h) regarding the rationale for "adjusting" ACC scores for AS (see Appendix H) seemed equally applicable regarding the use of any "intermediary key". Still another possible source of bias concerned the use of Q-statements based on Hurray's system of needs. Since these items may have been more readily translatable into "diag- nostic hypotheses", this could have made prediction from the interview sessions more difficult. It would have been possible to use items taken from ther— apy interviews for the Q-statements~-e.g., the trait pOpulation devised by Butler & Haigh (85)—-but this may have biased the results in the other direction. Obviously, Q-items derived from both sources would have resolved this difficulty, but such a pepulation of statements was not available to the writer. The decision to use the statements which were employed in this study was based on the following considerations: (a) the statements were available to the writer; (b) the statements -64- had been scaled for social desirability; (c) bias in favor of the "diagnostic prediction" was seen as less damaging. The rationale for reason (0) was as follows: The Judges were asked to predict the self-sort of a client. If one thinks in terms of "levels of consciousness”, the material contained in the interviews is "closer" to the level at which 0 would Operate in doing his own self-sorts; the level at which the diagnostic material provides information about 0 is much further-~"deeper", if you will—-from 0's sorting level. Consequently, there might be some bias in favor of more ac- curate prediction based on the interview material inherent in the experimental design. By using items which may tend to favor prediction fr m the diagnostic material, the intent was to cancel out some of this "interview-bias". Whether or not this decision achieved the desired result cannot be adequately determined; therefore, the origins source of the trait pOp- ulation remains as a bias in this investigation. The limitations discussed in this section require that interpretations of this investigation be made with caution. ~65- GENERAL IMPLICATIONS AND SUGGESTED FURTHER RESEARCH The major finding of this study was the decrease in predictive accuracy which occurred when the Judges were re- quired to make a second prediction of 0's self-sort (this "was particularly evident when the two predictions were based on different types of stimulus material). This result is a confirmation of findings in other studies (20,5%) and it is quite provocative.1 It seems-to contradict the truism: the more you know about a person the better you understand him. The data ofthis study suggests the "more" in the above adage requires further elaboration. One might raise the point whether the "more'information" implied above should be read "more of the E322 king of information". That is, hearing ten therapy interviews might lead to more accurate prediction than listening to one or two sessions; similarly, predicting from a battery of ten projective tests might lead to more ac— curacy than predictions based on one or two tests. These consideration could be subjected to an empirical test by requiring Judges to make a succession of predictions (at least more than the number required in the present study) 1 One interesting area of applicability would be the results obtained by Kelley & Fiske (60). -66... while the amount and kind of information provided was con- trolled. If the above speculations are valid, the successive predictions when the type of material is keptthe same (all interview or all diagnostic) should become increasingly more accurate; whereas, successive predictions based on, say, the following sequence I1,D1,12,D2,I3,D3, etc., should result in increasingly poorer predictive accuracy. The fact that for all Judges the second prediction was made with the same stétéfigfitgzgfiggzggé several factors which may be contributing to the obtained decrement in the second prediction session. Reusing the same items under different instructions could be a determining factor. This explanation, while pertinent to the present study, would not be supported by the results obtained in those investigations in which the same population of statements was used for all the sortings. Unless one is willing to posit a cumulative effect in which each successive sort becomes increasingly inaccurate due to the effects of the preceeding sorts, such an explanation does not seem tenable. It would be possible to test the appropriateness of the above position with the following experimental design: Two groups of Judges attempt.to predict the self-sort of an 0 from the same stimulus material. Group 1 uses the same Q- statements for all the sortings (self- and predictive-sorts); Group 2 uses a different trait-population for each sorting. -57- If there is a "building—up of interference" then Group 1 should show a decrement on the second prediction and Group 2 should not show this effect. Another possible consequence of using the same items for both the self—sort and the second predictive-sort in- volves the issue of threat. It may be that a Judge who "sees" the O as being severely disturbed, attempts to sort the items on the second predictive task differently than he sorted these same items for himself. (The fact that 7 of the 16 J5 who had an RS significantly different from zero at the .05 level or better, did not have a decrease in their ACC scores on the second prediction, tends to vitiate this arguement.) There is still another possibility, however, stemming from the fact that there tends to be a fairly high correlation between the self-sorts of the Judges and that of the O (median BS is +.M39). If a Judge was trying to predict O as being different from himself, then, conceivably, his predictive accuracy might diminish. Two methodological "wrinkles" if added to the present experimental design, would permit an empirical assessment of the above speculation. {eeping all other procedures the same, half of the Js in each experimental condition could have been told that O was a patient in an out-patient clinic; the other half of the Judges in each condition could be told that O was a student. If the factor which is responsible for the de- -68- crease in accuracy is the threat involved in predicting a disturbed person as similar to oneself, then the effect should be greater in the first group. A more subtle procedure would be to have half of the Js do their self-sorts first (as in this study) and half of them do the self-sort after the second prediction. If RS for the first group is consistently higher than for the second one, the importance of threat as a variable would have been demonstrated. That is, the second group of Judges would have systematically attempted to "see" themselves as different than the 0. It would be possible to combine the two above mentioned procedures into one experimental design. This would permit the measurement of the effectiveness of the different in- structions. Presumably, this could result in a measure of credulity; a factor possibly influencing the obtained results. A final factor which may be contributing to the decrement in ACC from session 'one' to session 'two', and which has im— portant (though rarely emphasized) implications for diag- nostic and psychotherapeutic theory, is the role played by feed-back.2 In the course of this investigation it became clear, particularly in the I"D and D I" conditions, that the Judges 2“The implications of the role played by feed-back in regard to diagnosis and therapy were pointed out by Dr. D.L.Grummon. ~69- were'weighing” the infOrmation acquired from the two stim- ulus sources with which they were provided. Several Judges, for example, commented under the I"D condition: ”The tests make him look 'sicker' than I thought he was from the inter- view material"; when I" followed D, a Judge might say: "I guess he's in better shape than I thought he was from the tests". These comments indicate that the second stimulus material was being evaluated in terms of the Judge's initial impres- sions of O (gained from the first stimulus material). Yet the Judges‘were unaware of the accuracy of their impressions; of the 16 J5 whose first nCC score was significantly different from zero at the .05 level or better, only two reached this level of significance on the second prediction. The Judges' inability to assess how accurate their impressions were could be a factor in the decrease in predictive accuracy. On the basis of the Judges' comments, it would appear that the feed-back they received-~the information contained in the second stimulus material--was misinterpreted. Almost invariably a Judge would ”correct" his initial ideas in terms of the cues he perceived in the second stimulus. In this ef- fort to gain greater accuracy, the Judge revised his initial impressions (which unknown to him were relatively accurate) and, as a consequence, he became less accurate. It would appear, then, that the Judges in this study were operating with the erroneous conviction that "more in- -70- formation" was equivalent to "more accurate information". In highlighting the role played by feed-back in this study, it is felt that perhaps a key concept in the differ- entiation between a diagnostic and a therapeutic approach has been brought into focus. In therapy sessions one has con— tinuous feed—back; in diagnostic testing one does not.3 The feed-back in therapy enables the therapist to continuously correct his impressions of the client. In this way the ther- apist attains an ever increasing, more accurate understanding of his client's thoughts, feelings, loves, hates, etc. Diagnosis, on the other hand, is a system of one-way communication. The diagnostician has almost no feed-back on the impressions he obtains from the diagnostic material. Thus it would seem, that feed-back in the therapy inter- views provides the therapist with a type of learning situation. That is, the therapist is able to test his hypotheses regarding the client continuously, retaining those which receive confirm— ation-- depending on the theoretical orientation of the ther- apist-~and discarding those which do not. One wonders whether approaching the therapeutic interaction in these terms does not provide an avenue for the application of many of the con- 03 learn», {MW ceptsAto this area of clinical psychology. 3 Bruner's formulation assumes a continuous feed—back process. Since in this study this condition did not obtain the results of this investigation cannot be considered a crucial test of his position. Rather as stated previously, Bruner's theories served as the theoret cal framework for this research. -71- iv . {Whig ‘jw The concept of feed-back would appear to have still a further application to Bruner's formulations. It is suggested that the process which Bruner terms ”a closer look" is, in essence, feed-back; while gating results from no feed-back or the ”wrong kind" of feedeback. In other words, where the feed-back and the individual's categorization system are con- gruent——of the same kind or at the same level of awareness-- the result is more veridical perception as in the "closer look". However, when there is incongruence between feed-back and caté egorization system, gating results and the ability to profit from additional stimulus cues is diminished. To translate these Speculations into the terms of the present study, the experimental design used provided the Js with only a limited amount of feed-back. The Judgeswere able to assess their impressions in terms of additional in- formation, but they did not have any objective criterion on which to based this assessment. In the heterogeneous con- ditions (I"D and D I") the second stimulus material--the ad- ditional material or feed-back—~was different,than that used by the Judges to gain their initial impressions. Thus, in the heterogeneous experimental conditions an incongruence existed between the initial categorization system and the feed-back available to the Judges. As a consequence, the feed-back appears to have been misinterpreted; the Judges altered their initial impressions resulting in less accuracy. -72- . '. ltd-'JMKJ’ w u ‘ In view of these speculations (and the results obtained in this study) it would appear that the usefulness of pre- therapy formulations is Open to serious questioning. The data appear to lend support to Rogers' (87) reluctant con- clusion that diagnosis as a necessary prelude to psycho- therapy may be a ”protective alternative to the admission that it is, for the most part, a colossal waste of time." It may well be that the vast amount of time spent, prior to beginning a psychotherapeutic relationship, in formulating plans, goals, dynamic characterological descriptions, etc., could be more fruitfully utilized., The results of this inveStigation have implications not only directly related to the specific issues to which the study was addressed, but ramifications in other areas as well. This research employed a methodology which might easily lend itself as a method for pairing therapists and clients. Given a recorded intake interview and the self-sort of a prospective client, the selection of a therapist could be made on the basis of which of the available therapists could most accurately predict this individual's self-sort. Indeed, these ACC scores could be the independent variable and the success of the therapy the dependent variable in such an inVestigation. The ACC scores could be varied, unknown to the therapists, through a careful pairing of therapist and O. -73- T Although it was not the major focus of this study, the failure to replicate the clearcut relationship between level of experience and accuracy of prediction reported by Kell (59), suggests that a more intense investigation of experience as a mediating variable in this type of research should be done. The methodological innovation introduced in this study of using two equivalent populations of Q-statements (containing clusters of items) appears to be justified; particularly in regard to the increase in experimental control it engenders. The data of this study could be utilized in the development of a well standardized, reliable, and internally consistant (insofar as the clusters are concerned) trait population for use in future research involving Q—methodology. Through an analysis of the self-sorts of the clinicians, hypotheses for additional research could be generated. For example: what are the characteristics of the "good predictor"?; What is the realtionship between "diagnostic" or "empathic" orientation and experience?; Which orientation is associated with a higher degree of successful outcomes in therapy? The general focus of these suggested research projects would be an attempt to isolate factors, both in the therapist and in the therapeutic relationship, which would be highly related with successful psychotherapeutic interactions. -74. SUMMARY AND CONCLUSIONS This investigation attempted to provide empirical data pertinent to both a practical problem and a theoretical form- ulation. The former concerned the usefulness of formal diag- nostic procedures in therapy; the latter dealt with a hypo--~ thetical construct--gating-—introduced bleruner in a theory relating to perceptual phenomena. To the extent that the ability to predict a client's self-sort--being aware of a client's phenomenological world --is deemed an important aspect of psychotherapy; and, to the extent that Q-methodology provides an adequate, quanti- fiable method for ascertaining a client's self-concept, the results of this study appear to have pertinence for the above mentioned practical problem. The data in this sttdy fail to confirm either the "diag— nostic" or the ”a-diagnostic" viewpoints. It was found that utilization of formal diagnostic material neither aids nor hampers the ability to accurately predict a client's self- sort. That is, clinicians predict equally well whether or not diagnostic material is made available to them. However, if bgth diagnostic material and interview material are pro- vided, and predictions are made on the basis of both types of material (regardless of the sequence of presentation) clinicians tend to predict worse on the second occasion. -75- These results were interpreted as indicating that the categorization system the Judges develOp (based on the first stimulus material) "interferes" with the full utilization of the information provided by the second stimulus material. The above ste tement is, in essence, a rewording of the gating pr nciple which provided the model upon wlu this study was based. he general prOposition states: the type of preconceived categorization system.with which one approaches events will affect the sensitivity to nuances in those events Support for the validity of the proposition would obtain from the verificat ion of two mo: e specific hypotheses. Iypoth- esis 1 stated: Mi sensitivity of an individual to nuances of a situation will be diminished if he ap- proaches that Situation with a preconceived categorization system In the main, the da a of this study supported this con- + tention. It was found that, regardless of the type of mater- ial employed or the sequence of presentation utilized, a marked tendency exiSted for clinicians to predict signifi- cantly poorer on the second of two prediction sessions. Hypothesis 2 stated: the sensitivity of one individual for another per son will increase as the degree of con- gruence between the categorization systems of the two individuals increases ~76- The results obtained were in the general direction of supporting this hypothesis; however, they were not at the level of statistical significance. Consequently, the findings of this investigation offer only partial support for the general proposition. Signifi- cant evidence for the existence of a type of gating mechanism' was found. :r an attempt to discover a factor which might' mediate the functioning of a gating process--the existence of a high degree of similarity between the persons involved in the interaction-~the evidence is merely suggestive. -77... 1O 11 12 13 B I B L I O G R A P H Y Alexander,F & French,T.M., gt al Psychoanalytic Therapy. Ronald; N.Y. 1956 Allport,F.H. Theories g; Perception and the Concept Structure. Wiley; N.Y. 1955—7 Ash,P. "The Reliabilit of Psychiatric Diagnosis" g.Abn. é Soc. 19H9 3 :272-277 Baker,B.O. & Sarbin,T.R. "Differential Mediation of Social Perception as a Correlate of Social Adjustment" Sociometry 1956 xix:69-83 Bellak,L., Pasquerelli,B.A., & Braverman,S. "The Use of the TAT in Psychotherapy" g.Nerv. g Ment. Dis. 19H9 110:51-65 Berg,I. "Response Bias and Personality: The Deviation Hypothesis" {i.Psychol. 1955 HO:61-72 Bender,I.E., & Hastorf A.H. "0n Measuring Gereralized Empathic Ability (Social Sensitivity)" g,§pg. §_§gg. 1953 h8:503-6 Blake,R.R & Vanderplas,J.M.. "The Effect of Pre- recognition Hypotheses on Veridical Recognition Thresholds in.Auditory Perception" g.Person. 1950 xixz95-115 A Block J.A. "A Comparison of the Forced and Unforced -sorting Prodedures" g.§d.Psych.Meas. 1956 16: #81 -1+93 Bordin,E.S. "Diagnosis in Counseling and Psychotherapy" g,§Q.Psych.Meas. 19%6 6:169-18H Bruner,J.S. "On Perceptual Readiness" Psych.Rev. 1957 6hz123-152 Bruner,J.S., Goodnow,J.J., & Austin G.A. A_Study _o_f_ Thinking. Wiley; N.Y. 195% Bruner,J.S. & Postman,L. "On the Perception of In- congruity” pg.Pers. 19h9 xviii:206 .73.. 1a 15 16 17 18 19 2O 21 22 23 2% 25 26 27 28 Brunswik,E. Systematic and Re resentative Desi n of Psychological .Eggoeriments. . alif. ress;BerEeI'y' 1949* Cameron,D.E. "A Theory of Diagnosis" in Current Prob- lems ings chiatric Diagnosis. (Eds. Hoch,P.H. & Zubin,J.7_%rune &“Stratton;‘N.Y. 1953 Cohen,J. "An Aid in the Computation of Correlations Based on Q-sorts" Psych. Bull. 1957 5hz138-9 Coleman, J.L. "Teaching Basic Psychotherapy" Amer. g. Ortho. 19%? 17:622-7 Cronbach,L.J. "Correlation: Between Persons as a Research Tool" in Ps chothera : Theor and Re- search. (Ed. MowrWer, . . onald; Ix. . 7953" Cronbach,L.J. & G1eser,G.G. "Assessing Similarities Between Profiles"' Psych.Bull. 1953 50:h56-#72 Dailey,C.A. "The Effects of Premature Conclusion upon the Acquisition of Understanding a Person" g.Psych. 1952 333133-152 Deabler,R.F. "The Therapeutic Use of the TAT" ‘§.Clin. Psych. 19h? iii:2h6-252 Duncan D.B. "multiple Range and Multiple §_Tests" Biometricg 1955 11:1-h2 Dunnett,C.w. "A Multiple Comparison Procedure for Comparing Several Treatments with a Control" g.Am . Stat.Assoc. 1955 50:1096-1121 Dymond,R.F. "A Scale for the Measurement of Empathic Ability" pg.Consult.Psych. 19%9 13:127-133 Dymond R.F. "Personality and Empathy" ‘g.Consult.Psych. 1650 male-350 m Dymond R.F. "Can Clinicians Predict Behavior?" g.Pers. 1553 xxiiz151-161 Ebel,R.L. "Estimation of Reliability of Ratings" Psychometrika 1951 16:407-h2h Edwards A.L. Ex erimental Design in PsychOIOgical Research ."JT—R nan-”art ;"N .Y . 1 956 -79- 29 30 31 32 33 3% 35 36 37 38 39 #0 Ln Edwards, A. L. "The Relationship between the Judged Desirability of a Trait and the Probability the Trait will be ndorsed" J.Appl.Psychol. 1953 37:90-3 Edwards,A. L. Personal Preference Schedule; COpyright: 1953 Psych. Corp., N.Y. Edwards,A. L. & Horst, P. "Social Desirability as a variable in Q Technique Studies" J. Ed. Psych. Meas. 1953 13: 620-5 Fenichel, 0. Problems of Ps choanal tic Tec Migu . Psychoanalytic fiQuarterly; N. Y. 19 1 Fiedler,F. E. "A Method for Objective Quantification of Certain Counter-transference.Attitudes" g, Clin. Psych. 1951 vii: 101- 7 Fiedler,F. E. "An Exploratory Study of Unconscious Feeling Reactions in 15 Patient-Therapist Pairs" J .Abn. & S__9__c. 1952 117:19-16-1153 Fiedler,F. E. "The Psychological-Distance Dimension in Interpersonal Relationships" J. Pers. 1953 xxii:11+2-150 Fiedler,F. E. "Quantitative Studies on the Role of Therapist's Feelings towards their Patients" in Psychotherapy:Theor and fiesearch. (Ed. Mowrer, O. H.) Ronald; N.Y. 953 Fordyce,w.E. ”Application of a Scale of Dependency to Concepts of Self, Ideal-Self, Mother, and Father" Unpubl. PhD Thesis, U. of Washington; 1953 Forer,B.R. & Tolman,R.S. "Some Characteristics of Clinical Judgment" .J.Consult.Psych. 1952 16:3H7-352 Frank,G.H. "Note on the Reliability of Q-sort Data" Psychol.R ept . 1956 2: 182 Frank,L.K. "Projective Methods for the Study of Personality" J. Psychol. 19398 :389-512 FrommpReichmann,F. Principles of Intensive Psychothera py U.Chicago Press;‘§h1cago. 19 O ..80... 1+2 #3 #5 #6 1+7 #8 #9 So 51 52 53 Gage,N.L. "Judging Interests from Expressive Behavior" Psych.Monogr. 1952 66:#18 Gage,N.L. "Accuracy of Social Perception and Effective- ness in Interpersonal Relationships" J,Pers. 1953 xxii:128-141 Gage,N.L. & Exline,R.V. "Social Perception and Effect- iveness in DiscusSinn Groups" Hum.Relat. 1953 6:381-396 Gage,N.L. & Chronbach,L.J. "Conceptual and Method- ological Problems in Interpersonal Perception" Psych.Rev. 1955 62:H11-%20 Gage,N.L., Leavitt,G.S., & Stone,G.C. "The Inter- mediary Key in the.Analysis of Inter ersonal Perception" Psych.Bull. 1956 53:25 -266 Gibby,R.G. "The Stability of Certain Rorschach variables Under Conditions of Experimentally In- duced Sets: I The Intellectual Variables" '1. Proj.Tech. 1951 15:3—26 Gilbert W.H. "Counseling: Therapy and Diagnosis” pp.351- 378 in Annual Review 9; Psychology: 1952 ; Annual Reviews, Inc.;Calif. . Gill,M., Newman,R., & Redlich,F.C. The initial Interview_in Psychiatric Practice. International Universities Press; N.Y. .1955 I Gollins,E.S. "Forming Impressions of Personality" g. Pers. xxiii:65—7 Goodnow,JJ.& Pettigrew,T.E. "Some Difficulties in Learning a Simple Pattern of Events" Paper pre- sented at Annual Meeting of Eastern Psychological Assoc. at Atlantic City; March 1956. Quoted in ( Halpern,H.M. ”Empathy, Similarity, and Self-Satis— faction" ‘J.Consult.Psych. 1955 19:h#9-#52 Hastorf,A.H. & Bender,I.E. "A Caution Respecting the Measurement of Empathic Ability" J,Abnfl4g Soc. 1952 #7z57H-6 A . . s . q . . . . . . a u n a . . u n A a o ‘ . ¢ 1 . . n . c . . . y. « . . .. . . J . u . a u . . - o . . o . . a an 55 56 57 58 59 60 61 62 63 6h 65 66 67 Hathaway,S.R. "Clinical Intuition and Inferential Accuracy" ‘l.Pers. 1956 xxiv;223-250 Henle,H. "An EXperimental Investigation of Past Experience as a Determinant of Visual Form Perception" g.Exp.Psychol. 19H2 30:1-22 Herrock,J.E. & Nagy,G. "The Relationship Between Ability to Make a Diagnosis and to Select Appro- priate Remedial Procedures" g,Gen.Psychol. 19H8 Johnson,D.M. The Psychology 9; Thought and gudgment. Harpers; N.Y. 1953 . Jones E.E. & de Charms,R. "Changes in Social Per- ception as a Function of the Personal Relevance of Behavior" Sociometry 1957 20:75-85 Kell,B.L. "An Experimental Study of the Ability to Predict the Self Concept of an Individual from his Therapeutic Behavior" Unpubl. PhD Thesis; U.Of Chicago: 1951 Kelley.E.L. & Fiske,D.w. The Prediction of Per- formang§_ig Clinical PsychoIogy. U.of KIcEIgan Press; Ann Arbor 1951 Kelly,G.A. Ps cholo ‘2; Personal Constructs:Vol. I. Norton; N.Y. 1955 King,F.w. "The Use of Drawing of the Human Figure as an Adjunct to Psychotherapy" ‘1.Clin. Psychol. 195% 1:65-9 Kleet,C.J. "The Stability of the Social Desira- bility Scale Values in the Edwards Personal Pref- erence Schedule" g,§pnsult,Psychol. 1957 21:183-5 Livsen,H.H. & Nicholes,T.F. "Discrimination and Reliability in Q-sort Personality Description" g,Abn.§_Soc. 1956 51:159-165 Luchins,A.S. Mechanization in Problem Solving. Psych. Monographs 19HZIEWT¥2HBI Luft,J. "Implicit Hypotheses and Clinical Prediction" g,Abn. Q Soc. 1950 %§:756-9 Luft.J. "Differences in Prediction Based on Hearing vs. Reading Verbatum Clinical Interviews" 1. Consult.Psychol. 1951 15:115-9 -82- 68 69 7O 71 72 73 7% 75 76 77 78 79 80 81 Lyerly,S.B. "The Average Spearman Rank Correlation Coefficient" Psychometrika 1952 17zh21-8 McClelland,D.C., Atkinson,J.W., & Clark,R.A. "The Projective Expression of Needszlll. The Effect of Ego-Involvment, Success, and Failure on Perception" g. Psychol. 19%9 27:311-330 Mehlman,B. "The Reliability of Psgchiatric Diag- nosis" g.Abn. a Soc. 1952 7:577- Menninger,K.A. ‘A Manual for Psychiatric-Case Stu z. Grune & Stratton; N.Y. 1952 Meyer,M.M. "The Direct Use of Projective Techniques in Psychotherapy" g,Prol.Tech. 1951 15:263-7 Muench,G.A. "The Application of Diagnostic Methods to Counseling and Psychotherapy" pp.53-72 in Recent Advances in Dia nostic Testin . (Ed. Harrower, M.R.7”Thomas; SpringfieId, Ill. 19 O Murphy,G. "Affect and Perceptual Learning” Psych. Rev. 1956 63:1-15 Murray,H.A. Egplorations in Personality. Oxford University Press; N.Y. T938 Notcutt,B. & Silva,A.L.M. "Knowledge of Other PeOple" Patterson,C.H. "Is Psychotherapy Dependent upon Diagnosis?" Amer.Psychol. 19H8 3:155-7 - Postman,L. & Bruner,J.S., "Perception Under Stress" Psych.Rev. 19MB 55:31H-323 Postman,L., Bruner, J.S., & Walk,R.R. "The Perception of Error" Brit.g.Pszchol. 1951 p.1 ff. Powdermaker,F., "Techniques for the Initial Inter— view; Methods of Teaching Therapy" Amer.g,Psychia. 1998 10%:6h2-6 Quinn,R.D. "Psychotherapists' Expressions as an Index to the Quality of Early Therapeutic Relation- ships Established by Representatives of the Non- directive, Adlerian, and Psychoanalytic Schools" Unpubl.PhD Thesis; U.of Chicago 1950. Quoted in (37) ~83- 82 83 84 85 86 87 88 89 90 91 92 93 9h 95 96 Rapaport, D. "Projective Techniques and the Theory of Thinking" pp. 196-203 in Ps choanal sis, Psychiatry, and Psychology. (ed. R. R.Knight5 International Univer- sity Press; N. Y. 195% Remmers, H.H. "A Quantitative Index of Socio-Psychological Eupathy" Amer.J.0rtho. 1950 20:161-5 Rogers, C.R. Client—Centered Therapy. Houghton-Mifflin; Boston. 1951 Rogers, C.R. & Dymond,R. F. Psychptherapy and Personalit Change. U. of L’hicago Press;vhicago. 195$ Rogers,C.R. "The Case of Mrs. Oak" in (85). Rogers, C. R. "The Necessary an Sufficient Conditions of Therapeutic Personality hange" IJ.Consult.Psychol. 1957 21: 95-103 "' ""y . Rogers,C.R. "Significant Aspects of Client-Centered Therapy” Amer.Psychol. 1996 1 :h15—922 Saul, L. J. "The Psychoanalytic Diagnostic Interview" Psa.Q. 1957 xxvi: 76-90 Schafer,R. Psychoanalytic Interpretation of Rarschach Testing. Grune & Stratton; N. Y. . 195E Siegal, S. Nonparametric Statistics for the Behavioral Sciences. McGraw-Hill; N .Y. 1956 Siipola,E.M. AG Grou up Study of Some Effects of Pre- paratory Set. Psych.Monogr. 1935 E6:%210” Soskin,w. S. "Bias in Postdiction from Projective Tests" J. Abn. & Soc. 195% #9z69-7H Stanley,J.C. "Statistical Analysis of Scores from Counter-balanced Tests" _.E§E.Ed. 1955 xxiii: 187-207 Stephenson,w. The Study 9; Behavior. U.of Chicago Press; Chicago. 1953 Stone,G. C., Leavitt, G. S., and Gage, N. L. "Generalization of Accuracy in Predicting Standard Persons" Unpubl. Manuscript, Bureau of Educational Research, U.of Ill. 1959; quoted in (#5). .gml. 97 98 99 100 101 102 103 101+ 105 106 107 108 109 Taft,R. "The Ability to Judge People" Psych.Bull. 1955 21:1-23 Tagiuri,R., Blake,R.R., And Bruner,J.S. "Some Determinents of the Perception of Positive and Negative Feelings in Others" _J_._A_‘p_n_. §c_ _S__o_<_:_. 1953 h8:585-592 . . Thorne,F.C. "Directive Psychotherapy: IV. The Therapeutic Implications of the Case History" ‘1.Clin.Psychol. 1995 1:310-330 Thorne,E.C. "The Clinical Method in Science" Amer.Psychol. 19H7 2:159-155 Thorndike,R.L. Personnel Selection. Wiley; N.Y. 19H9 Thornton,G.R. "The Significance of Rank Difference goefficients of Correlation" Psychometrika 1983 :211-222 ' . 1 A Tryon,R.C. "Reliability and Behavior Domain Validity! Reformulation and Historical Critique" Psych.Bull. 1957 593227-299 Tupes E.C. "An Evaluation of Personalitberait Ratings Obtained by Unstructured Assessment Interviews" Psych.Monogr. 1950 6#:#317 Vernon,M.D. "The Function of Schemata in Perceiving" Psych.Rev. 1955 62:180-192 Watson,R.I. The Clinical Method lg Psychology. Harpers; N.Y. 1951 . . _ Wert,J.E., Neidt, 0.0., & Ahmann,J.S. Statistical Methods 13 Educational and Ps cholo ical Research. Kppleton-Century—Crofts, Inc.; N.Y. 195% Wolf,R., & Murray, H.A. "An Experiment in Judging Per- sonality" J.Psychol. 1936 3:358 Wyatt,D.F. & Campbell,D.T. "On the Liability of Stereotypes of Hypotheses" J,Abn.‘g_Soc. 1951 #6:h96-500 . _ A ***** *** 3k -8 5... APPENDIX A LIST OF JUDGES Behrle, Fredrick J. m. Fredrick B. Brice, Norman Bonier, Richard S. Brownfain, John J. Cain, Albert Capell, Martin D. Chodorkoff , Bernard Erlandson, Forrest L. Freed, Griffith o. Feuerfile, David F. Gentry, George M. Grater, Harry Hyman, Marvin Lesser, william Macs, John ~86.- Males, Herbert B. Margolis, Marvin. McDonough, Joseph M. Mikel, Bernard Monteiro, Antonio Papania, Ned Pilisuk, Marc Heisman, John. M. Roberts, Allyn F. Silver, Albert Silverman , Herbert Smith, Walter G. Stein, Jacob Vidulich, Robert N. Vogtman, Walter G. Wehl, Julian APPENDIXLBs EQUAIING PROCEDURES EUR,SOCIAL.DESIRABIEITI’ Edwards' Personal Preference Schedule is constructed in the following manner. ...the pairs.of statements comprising the items of the PPS have been matched with respect to their social.desirability scale values, in as far as this was possible. The statements were sealed for social desirability using the psycho- logicalpscaling methods of successive intervals.... When statements from two differing variables ‘were paired to make up an item in the PPS, the statements were selected so that the social de- sirability scale values would be comparable. The extent to which this matching was successful is indicated by the intraclass correlation of .85 between the social desirability scale values of the paired statements making up the items. It was anticipated that matching statements for different personality variables.upon the basis of the social desirability scale values of the statements would tend to minimize social desire. bility as a.factor influencing responses to the items. (30) The statements used in this test are based on.Mhrray's 15 need categories. For each need eight.or nine statements of presumably differing social desirability value were em- ployed. Every need.categoryh-i.e., the statements represent- ative of these need categories-ois paired with every other need category twice; once coming first and once presented second in the pair. The subject is required to subscribe to one statement in each pair as being true about him. Through careful selection it was possible to assign ~87.- statements representative of the different need. categories to the two. Q—arrays and still preserve the balance in social desirability» that had been achieved in the. construction of the test. For example, Item 6 of the PPS contrasts a statement in- dicative of i Achievement with a statement representative of ii Deference; Item 11 contrasts the same two need. categories but in the reverse order and with different statements. fl Achievement (Item 6). and E Defermce (Item 11) went into Q"; i Achievement (Item 11) and ii Deference.(Item.6)‘ went into Q'. This "trading off“ of statements. continued until each need category was represented by four items. in each Q-arrey. ~88... APPENDIX.C SUGGESTED INSTRUCTIONS FOR Q-SORT Step 1: Remove the cards numbered 1 to 9 and put them to one side. Step 2:. Read the first statement. If it is descriptive of you, put it on the table at the right. If it is not descriptive of you put it to the left on the table."lf you are doubtful or not sure that the statement is descript- ive of you put It directly in front of you; Step 3:: Read each statement in turn, putting them into one of the.three piles. To the right you will have all.the statements that you think are descriptive of you. Directly in front of you will be those statements about which:you are doubtful. To your left will be all those state- ments that you think are not descriptive of you. Step he Now arrange the cards numbered 1 to 9'in front of you with the number 1 card to the extreme left and the number 9 card to the extreme right. Think of these cards as representing a.sca1e with the "1" card meaning the LEAST and the "9" card meaning the MOST descriptive of you. Step 5: Take the pile of statements at your extreme left and select the one statement that is LEAST DE» SCRIPTIVE of you. Place that statement on top of card nuns ber1. Step 6:. Take the pile of statements at your extreme right and select the one statement that is BEST DE; scalgrm of you. Place 13551: statement on tap of card num- ber 0 Step 7: Repeat Step 5 for card number 2, using the state- ments in the left pile. Do the same for card hump bar 8, using the statements in the right pile. Note that you are to put three statements on each of these cards. Step 8: Continue in the way described above working from the extremes toward the middle tilw you have ex- actly the required number of statements n each tpile as that printed on the numberedm 11 f sort the three middle pilesdfifi, 32231 you;%%rt ea§ier t0 the extremes to the middle. -89.. Eote: If you use up all the statements in either of the initial three piles before you have completed all the sorting then select statements. from the middle or doubtful pile, to complete the exact number require by each; Card. . Ion may find it useful and convenient to put the one above the other for each of the numbered cards. In. this way you can shift statements back and forth from one card to another. You may make any changes you desire. Just be sure that you have the proper number of statements on each card. Step 9:. When you have finiShed, put the statements for each numbered card under that card. Pick up the nine pile of statements, putting the number 1 pile on. top and the number 9 pile on the bottom. Then replace the rub- ber band on the pack. ~99- APPENDIX.D STATEMENTS OF Q‘ AND Q" ARRAYS Qt ABASEMENT I should confess the things that I have done WTQngo I am: timid in the presence of people I regard as my superiors. When things go wrong for me, I am more to blame than anyone else. I am.inferior in most respects. ACHIEVEMENT I would like to be a re- cognized authority in some Job, profession, or field of specialization. I am usually successful in things undertaken. I want to accomplish something of great signifi- cance. I enjoy doing tasks that others recognize as requiring skill and effort. AFFILIATION I participate in groups. I often write letters to my friends. I am loyal to my friends. I prefer to do things with my friends rather than by myself. “91... Q"? I feel guilty when- ever I have done something I know is wrong. I am depressed by my? inability to handle various situations. The pain and misery I have suffered has done me more good than harm. I feel better when I give in and avoid a fight. I can do things better than other people. I have usually done dif- ficult jobs.well. I would like to write a great novel or play. I like to solve puzzles that other people have dif- ficulty‘with. I have many friends. I share things with my' friends. I have strong attachments to people. I like to do things for people. Q! AGGRESSION I tell other people what I think of them. I like making fun of people that I regard as stupid. I attack points of view that are contrary to mine. I like to read newspaper accounts of murders. AUT ONOIAY I avoid situations where I an expected to do things in a conventional way. I like to do things my own way. I say what I think about things. I come and go as I want to. CHANGE I like to eat in new and strange resturants. I participate in new fads and fashions. I like to move about the country and live in different places. I like to do new and different things. DEFERENCE I like to praise someone I admire 0 When planning something, I get suggestions from peeple whose opinions I respect. In a group, I accept the leadership of someone else. -92- Q" I like telling people off when I disagree with them. I like criticizing some- one publicly. I get revenge when some- one has insulted me. I get so angry that I throw and break things. I like to avoid reSponsi- bilities and obligations. I like to be independent. I often criticize people who are in a position of au- thOTitYo I feel free to do what I want to do. I like to experience novel- ty and change in my daily rout- ine. I like to experiment with new things. I like to travel. I like to meet new peeple. I often tell my superiors they have done a good Job. I follow.instructions ° and do what is expected of me. I accept the leadership of people I admire. Q! I like to find out what great men have thought about problems in.which I an inter- ested. *DEPENDENCE I take pains not to incur the disapproval of others. I tend to be apologetic and don't stand up for what I know are my real feelings. I secretly wish that I were a child again. I feel things deeply and personally. I feel lost and helpless. DOMINANCE I like to be one of the leaders in the organizations and groups to which I belong. I argue for my point of View when it is attacked by Others. I like to be called up- on to settle arguments and dis- putes between others. When with a group of people, I make the decisions about what we are going to do. ENDURANCE I like to finish any job or task that I begin. I stick at a problem even when it may seem as if I am not getting anywhere with it. I work hard at any job I undertake. I like to put in long hours of work without being distracted. -93... Q it I like to read about the lives of great men. I give myself utterly to the happiness of someone I love. I do a great many things Just to avoid criticism. I feel afraid of being alone, of not being wanted. I have a strongu sense of resposibility. I feel "out of sorts" if I have to be by myself for any length of time. When serving on a com- mittee I like to be appoint- ed cha rman. I am able to persuade and influence others to do what I want. I like to supervise and direct the actions of other pe0ple. I am regarded by others as a leader. I like to complete a single job or task at a time before taking on others. When I have some assign- ment to do, I start in and keep working on it until it is com- pleted. I often stay up late in order to get a job done. I like to avoid being: interrupted while at my work. QI EXHIBITION ISM I often say things that are regarded as witty and clever by other people. I sometimes do things Just to see what effect it will have on others. I like to ask questions which I know no one will be able to answer. I like to talk about my achievements. HETEROSEXUALITI’ I like to become sexually excited. I engage in social activ~ ities with persons of the 0p- posite sex. I like to be in love with someone of the apposite sex. I often listen to or tell jokes in.which sex plays a maj- or part. *INDEPENDENCE I have considerable con- fidence in.my ability to do most of the things I undertake. I disregard regulations that hamper my freedom. I hesitate to consult any- one for help. I find it easy to turn down unreasonable requests. My acquaintances have a rather low standard of achiev- ment 0 ' INTRACEPTION I try to understand how people feel about various problems they have to face. -9)+.. Qt? I like to tell amusing stories and Jokes at parties. I tell other peeple about adventures and strange things that have happened to me. I like to use words that other peOple do not know the meaning of. I like to be the center of attention of a group. I am regarded as physical- ly attractive by those of the opposite sex. I often go out with et- tractive people of the Opposite sex. I like to kiss attractive persons of the opposite sex. I participate in discus- sions about sex and sexual activities. All modesty aside, I feel I am a pretty capable and comp- etent person. Frankly, I feel that I don't care to go out of my way for other people. I almost always try to avoid people's pity. I refuse to take sug~ gestions from others out of pride. I feel I have enough friends; if anything, too many. I often put myself in. someone“s place and to imp agine how I would feel in the same situation. Q! I think about the person~ alities of my friends and try to figure out what makes them as they are. I like to analyze my own motives and feelings. I like to analyze thee feelings.and motives of others. NURTURANCE I show a great deal of affection toward my friends. I treat peOple with kindness and sympathy. My friends confide in me and tell me their troubles. I like to do small favors for people. ORDERLINESS I have my life so arrang- ed that it runs smoothly and without much change in plans. I plan and organize any work that I undertake. I keep my things neat and orderly on my desk and work space. Any written work that I do is precise, neat , and well organized. SUCCORANCE I expect my friends to do many small favors for me cheer- flllly .- do \31 i C!" I like to predict how' my friends will act in var- ious situations. I judge people by why they do something-not by what they actually do. I like to study and to analyze the behavior of others. I readily forgive people who may sometimes hurt me. I like to help other people 0 I sympathize with peOple when they are hurt or sick. I am generous to my friends. I have my meals organ- ized and a definite time set aside for eating. I have my work organiz- ed and planned before beginn- ing it. I keep my letters and bills, and other papers neatly arranged and filed according to a system. I like to make plans before starting in to do some- thing difficult. I expect my friends to treat me kindly. Q ' Q" I like people to be I like pe0ple to symp- sympathetic and understanding athize with me and cheer me when I have problems. up when I am depressed. I like pe0ple to feel I like my friends to sorry for me when I am sick. make a fuss over me when I am hurt or sick. I like people to help me I like my friends to when I am in trouble. show a great deal of affec- tion toward me. .95 .. APPENDIX E RATING SCALE (ABASEMENT) (ACHIEVEMENT) (AFFILIATION) (AGGRESSION) fiiso' (ADTONOMY) (CHANGE) (DEFERENCE) (DOMNANCE) -97-' . . ex .1 . ' ' o r . V v m .. par. -.x - . , .- . .1..:‘ ‘ ‘. .‘ ' , ;. _‘ ~ . I ' | c I ' . w > , ‘ ‘ :' '4 1 - I . . - k. .,-. - .,_ -. ,‘>. - ..- . . ‘ . :1 . - . ._ . - . . . ,.. ,. . ‘ . «\ ... ‘ 1 ‘ < . .1 a. - -. -.‘. ..' .T . . ,- Z 4. y «‘4 , '4. ‘4 ... . .n .. . . 3 ._ .4 v ‘- ;-—v-A . -.. g as. '3. -1- -, ... inf. ' . _, .. .»-. a; :-_ .. :1 -' ., - -.. "' ) ‘-—-"‘.~~'. . .n 2.12;. -1 (j :41 - " 1 \ » “ -.P-u- .-. _ ‘,," -“ . ~R . g .. .. .. . o ‘ :— F . . .- .~ ‘ z .1. . . ' 1"-" >._.‘ ‘1'... . .-;‘ ‘ - _ "‘.“ '- . -- .. .1". ‘ ’-..',-‘ . '1‘4‘.".' "“‘,"1"":‘?‘ - .‘._-~. L.—k..( 3 ‘1 l‘ H -- . , ""‘. . 1‘44 "'12 .’---. ' 1"-.. -'1 ,-.- i. 1’ 1- ~' J m“ .' _‘,_,.,., "7 "J 1" .3.’_‘_ .31- _.\-$ ..l '- J 1 ~- ;. 1.“ .h' 9" f" ‘n :. x ‘ 1' . -. .. -i . u . , g . 6‘} ““1," . . I ‘ ‘1' I 4‘ u . -. .1 --. "'-""‘1 . - ”u‘. 5 ' 1 - A ‘ i P. I t"- (EmBITION) . ' 1 O. )'> _ ,_ . -. .-........- . . ~ ~ -. . .- - .— H .. ..r- . . ‘ '5". . .. a . m :"'-.;',2' 1"" 1* t r‘ ' -' -. "'1 . - '- 'xJ '3 '- .._.....'.. 5. a— 5" .1. ‘ r f. 24- -|. ‘ :':‘_ I.‘ J - ~1'.--q.a‘ '3 ‘- . '. L? [7 '.3 I 1 . . \- .. ‘ " " ‘ '4}- .‘p'hfl ’ . .. .jlv‘f‘l '._ (‘10? _,., u . ‘ . \ ' y ' ' ..1 _.cz.,..11-.,..1.' .. . .. .. 1:- ...-\.,¢’~ !. ,r', -. J ‘ s y , I - ’ ' ‘ ¢"“r‘ ‘ u ‘-\. .' 1. j" 1 - I . .)l - . l ‘.'o.-. ~11 \ . . .I . (\ ‘, 9y! . . ‘ ‘9 ‘...' .5 ;\ , . q . - . -. 3! ' ‘1 _ I -\ 7.... ..I , .— .. - \~od\ ~ p . fi ‘fi. . ‘O‘I . 1 " I’J -' .3. ‘ 11.‘ n'. . . . .)_. 1‘ :‘T‘ (‘1’. ‘f ph.\‘2"_‘3_ '3 it?‘ 5“ ' '~'\".‘\ _i‘ v . 1‘ 'rfi- -\.— -_-" ‘1». f\ " --« "7 is. ,-n (‘1 o I. - '.‘; I." - ~~5 f.‘;'/ 1. .._. \Q- -. M o. ' '.‘ .‘y'.l : .3 5. 1 "1 -‘.’i. l r} .1 "1.. ,- 1:-. .' \‘ . . E .‘ ‘..‘~ '.‘ l;- -« .~-'-~E-‘.o....- -..~' ._,. r-a '- \\_n~._,!'.; ~:-,1 ...'.«...-'_-.-I .lg-~':.4¢.v 1? 's- ‘4 .—..._~ ..I.\- _ .112. ‘. \..I9.; . t8 . __ v)"- 2 a. .- - .?._-‘, :u: - Jo. \ 1.". ; 'I ‘ ,’.v ,V ‘ ...y._.‘\ (A , . ~ "jl"' -‘ - .x '1‘ .'-' us 1 "135-. 4' ' .'_ . . .. v.3-“ . 4’ up ISL..-”.-" '_,'- ' ' (a . .~“'-'1‘" ‘ ,§‘- ‘. \ —“ ' "I g. . ‘_ ""' .. -,‘.‘ 5‘ i. v. I : ...‘_., I‘,’ _a 0’ '- - Jul- — —~ I.) 5. .. . t‘y - .. . ... .4 J! 4 .4 lV . -.,, r . , I. v. — ._ - .. e.— e._ ,‘F ‘,. 1..., . ‘ a... “I ¢ 1 as h .- H '- ) “ '. ' ' ‘ . ' ‘ " - .v - ' . '-\'- . -\ . . . \. - . ._ .'. - -,. p» -' .. . . A 1 - x. . ,- ’1‘ , . -_ _ - _ ‘. - . r '34}: .";,~,;‘!\ 1,. " ’~.Ur*";"': ,L - :-1;l_ gutx'L .3 .' "a ,.\ 5"..2 .-\ ‘N 1..11’w1“,‘." f ,"u" . - - -— . ~ 1 v mun-~4- .. .. -- a \.~ ~- -.. ‘ 1.‘ 1-~.» — J. IL. . - a. . v . -- -— - \ _'. ‘ ' 1 ‘ , ; 90"! l ‘ c .4 0" I‘_. “K3 I "lfl-o I}: R n r ’ C k u- Gui - D r . . i. ‘ . I .-__ 1. y u l "_ t. .. L} {W ‘- . ‘, ‘ “n" i '9 t f V‘ ",v' I, 'r" 1‘ ,1 v . —1 - - . find-Wu , Hunt: '1: .'. ‘ '. n -. III’T C 11. . p .jJ....J- LAD... '_:.-1 v ( [m E .1... \ )1») -g- -.< . :s—J' fr: -" . '* “fur-:1 or r! w 1. v;1; {9 :1 u' = - . J- . ‘ " .. (2‘. . “I vi 1.,» r 'l ;.‘ .g “i": , 1 a. 1‘33 . 2’11. grain? r:.'-r:-.ra-*s.' ..rr‘. «n .,, J. 2" ~-‘ 1- ‘1 K- ...g. ‘ u-l F11 I ' (A t" ”a L5 ‘ u .- 1.) I ' u: '51 LI (3 '7. an! {J x: .' 1' r -.. p ‘VLJ _I’ .1. ’~.—' 'r 'h v IJ' Q ~0 . ' f“ u .- - . a r. and .-. .P“ 1‘ ax 4 3 ‘w a. a “A “A. 1,.) I) 0‘.) ~ - ‘x 19", -*-- 1 l ‘1 fl ‘1 PW ' -- " 0") ' . . 1‘ .r - . 5 r'» ~ - . . 1 .‘ ' '1 -‘~/ “'1' .' .- .c 1;- , 4 y. 5' n— 1.. 2. (,3 v a). o ..— . ' s3... '1. 'u t..- 1 3‘ .za‘. . '.,.—\ ',"' I... " ..I 4 ". .. .v #5 _a. 1. f- - .1. I ' -.\ -. 2 at. 11» f- ’ ) x' 1, k r v 2'... ‘3 a i. .2 i -. It --v 7:5 a. .5:.7,) -!b ("'1‘ '-. "'1 (”x flé“ u'~. if. '. [1"nl-Lv11 ‘ “ -_ I. X . 'u,“ s g 5 u ' ~01..'-- Wu", o»!.;..-. 1 -4 .-. \ .‘u ._ 0:. ... . r—a’v. P: ,v~ ‘ \ —~ 9 5 a .‘ .. oi .CV' .:. i‘. in? 2]): “111.1 _1’. .o ...‘1.-,; ‘i. t? r. ar-l‘. .- 0"" t ". ’.‘ w': ._-‘ 1 {a . J. a. 1‘ , - a. '. I "F! », r...‘ if u I) f. K c. ,‘ t" 1!“ '-_‘_, r“ 0-; '. ‘f. I "..1 .I.~‘ ~‘.~ '1. ~ ‘1" ,' ‘ .1 — t ., .' '. { . ~ .a - £1. ,_., ._ ._ - ’1: 1i ._ r - . .- .-_, -.._ .. . . . ‘ 1 '1 v v. n 1. 2A . ,. fl y...\,.9, Var]. v. .4.» tr 9'," yl‘ .;|.I.J. » .‘.~. Y. «I» 5 -1116!" s .-__' 1):. ~.Lou'¢’”7 7 . r ‘ - ' v‘"- H :1. 3’ r" 5’; - z t I - .. {:l .. . . . ’1 \. 5-1- . 1. (J ‘a . v . ... . I 1'. ..g , ‘ n - .. . .- _. -r~. c"- ucap 5. 1.) ’—~“*; ; {I T," 3': 'JL-ri' 9 ”'1‘” 13 v I “we; .-' 3.}. .‘4L..§-..’./: -'-..‘t 1:-3” .‘...‘-.’. .21. 7"“! Q h, ' . ,3 g. 3‘, 9 ., - -'~ , a 1.. ~ - -1 . . . .. . ”f"; . f¢J‘J:’-?— ”—1 r‘ f’ ".';’. 7 .". ~.~'r' ,1 '5’ .g‘ '1} g {'3‘ «INC; ‘1'“. v». 1 .‘u: 4’. ~ .y -4241»... .:.. .25 i“, II... a} t.~~.. 1,. u....-..».‘.. r, .---. ,. t 31, “Iv-J: mu » ;. uq. . . 4 ,, I ' "’ 'ww- '7' z ".1 ~ C. - fl . ' ‘- -nou - l .. . I i .2! . :H‘ »'.‘-"1,’ “3‘1? 1f} ‘- $ {’1 1:. *‘. '~ 7’7...- ."-'5 : .1'1' - c.— h "‘ VJ us ,... ‘x-IU': ‘. ~94. - . ..1 Col-.1. 1. .- - ...~ -.' ..J ..z . e .- '.-.. -~ .1 “.1“ ~.-. .‘ x :J .K. J 3", '9 .. 1.. 1. 71'... .1. .1 f.» .. .. .. a , . ~ , . . .- ~ . -. . '1 .9 .- - s v- 5 ..1. "t V" a "‘. .0 1 I.‘ --.1") K" m"; r" ‘5‘" 3' ,‘ ' ‘15". .1" .". r_ 7". - .. “. A“ 5:1)"";)‘ 7“; h) _\-; ‘ " ‘ ‘ ‘ ' 1 R " P 3. if ' ‘3 x ' ' (1 I'd ’ " ‘ "- 4’. i1 2 0.. vl/Ib vs J-apptp ,— 9‘, -2 ~. ' .,.._, ) - -v . - . - p 1,: ...,.4 -.~~4,1.—~_. 1 w J . .1. . .1. .. t ..< 1. '. ...a “.5 ."-‘ m‘ y p. .'_. I ‘v ‘ . -.g q . ' .- a .. J . .- r .. v a... "“ ,- J '1' " '\ v» a‘ rev-“W ' ~‘ 2 '7" I If} (”51‘ ‘ ‘f - 1- .- (4-111 or utkr‘ )~ J,‘ '\)'. ' ' 1" $1 ennui—Kn -i1.~.4 (. -9... fun~ - “'." .1. .. -' w 1' a . ’4 try - ,-,¢,.. 1.1. 77'! 3'. w . ,." ”Jag. -- 13., .. ) m... L'- . :J . 'l‘ . - we. . n' rye 3‘ A} a ; a. ' I u:;-) .p' \j’ ~ "‘ ‘ ' ‘ L1; r -... -.3 1 - 1 -1' ,- - . 9": ca. .. . ....- _ ~ - . o u .- l—~< 1' WA ‘ V's/1*” ‘ {m " ’ "'7‘. i" . . 1~ , .1 . '. c"-..-54o .V a‘.",‘;— ~l\d.- ‘ 1:) L Q%‘:—." an.IA.- H." 3 -5. st. ' a 0‘ ' "‘ . '3'» "' Afifi A “‘4' ’~'~ 13”“. “\ "v ”13 ..1.. M1- 1 '-- \Jv' 1’ 140—“. a 9 ~~‘- 5‘ \‘gfle - .ughfi’ J . ‘1 1 ‘- UK a _ . 1 . _ I,_ '\ J’.‘ I, f“ 7,- \-.< 1" = '* “~ " ' “ 9 3 c; "-5 . ' \_ ‘ . V ‘a 0;...) ‘ 1./ r, "' " "n * -, "." «t I ’7" r . 1.— ." v .9.- "1» .)"§Kar~ I : ‘1' K. 1 ”I, '3 (I *7 H») 19 \ . 1‘” 1 ’ 11 V» ~' o J.... .. r... .--. a . .-.‘z 1 a)... 5- 1.0.- : .I :w: "9.. ’_, ‘7’ ‘0! LT; {..r ._.‘~_ ‘ . VIN..- F”; A“! V.” '17. O n ‘3" If.“ v“ 11 v ’- v4, , .‘a v12: - ‘ ‘ " .I u.v I“. n... ‘_ h ’{ 2"“ '8, ’I- ' ‘1 ' U :49. be 2.3:). mm‘fler 'N"-“ "m " I." «ma-u.- l «stadiums». - any" are». r-- «4mm. m1mr~i- ain't-Ma '-J-'.-¢'- r IF 5 A" -. ftp-.12. .. a- 14* ; 1»; .‘m-l n- acts-:4. a: . we ‘1m’:-:'.9"83. say—mun" ..II‘ _- 1 c. .1 .. v”. *_’. the prediction which ye t4, '5‘ -2 .31 P 4! g: :4... '1‘ U‘. ,.“:I~..f’} -‘ 1.. x. .98.. 1319 are.“ _1 2m? ...LJJ 1.1" f’keer-J\\. ~'6 E -71 '11: 42:)- (‘Nrg m8 4‘ — a I v...” at: a?» . - APPENDIX F EXPERIENCE DATA SHEET 1 J: 31.3.1211.é 319:! you :‘ 1-111211, In 5-411 ": '-.-1 “- 1121.113 (:35 a! 3' an): a and oi“ Such a: ma": 12mm Elsa-.59 indie-$23., 1'1; tuciz-fling 1.521113 rate yurg'mgf 0312 a. c:z:.é,723- Vein-.1"; F 2 .;sidc:'.~ yet-2'. min" as be ink, 0' m seating" (112...?) or mus, at the fillOI‘CJD." ' 1 . 9) hazzcaticisn {lb-..E‘quis't . l-z-e indicate your tiwr- peutic 0:522:12611 .......-~_.-.._......._..=........2, .....n.....-..2..‘.... 1 o Circle the app?" up; ”1 £7; a 1 mg cherspv- e mambo? 136101: 1116115222." ,, of ycu? :"e 21.192 (.3? 0111 mcsuit precarious 1.0 5313216110 Piaf. 171235156789 30 importance 132‘: Jane” myax ‘; £11715 Circle the number in front of the. statcuns‘z 73123.7»: 2.1211211 mos» assurately damn-fins ;mur fiL-IJZE'F"5.EL1EC3 wfiuh Qataori 110511)..2q1..1 761: head 0;? ‘v'=31'l“:39;0' 1 ’-'l 3'.) q 9 2’ haw 3721‘s. relati-wcfiy lit-’51: :11 "12;: ‘ ‘91: 31.111. -:;'..2 tbs-322::- e read extonai"'-.‘"-‘ on Qaeaz'izs but haves mew-r flan-a 35:2- pwmw f I ire-2'15 done one (.1? i=- ~11; 0113.2?1‘19‘. have done se'asral 0,31353'113 {mars than {1:90:10 0’: \A A.‘ I have. dune quit e a number of (fl—«aorta a. 1d ass-re: read a111,. .- of the lite 9:11am .11. Wis ‘acchciqueo 9.1: In tha 12;..pe“imout proper-2,, you 1,5131 be as “.-:cd ta 11'adic1. me Q:::- ,1 111' a patient or (1.. 1am. on fine basis of a moaaz‘dcd therapy hon 1, :12de on tea 1:1an of ding? osfic. natal" ”GL1 Give?» the number in ff-‘flt ’11 t}: Mutamu balmy which (135:. {1: ~21 your um. estimate of "he am: urncye of 3511:? p'esaitttlerza 1, 1 My predictive Qmaor't. based on the Miami.” material 1:111 be mare summit; than my prod-let- 1w,- Qaaort based on the diamontio watering 1 2 My preéiia‘l 1m chcrt based on the diagnostic 12115.13.me 12111 “In nor; acorn-cue than my prodiotivo (3,1180% band 0112 2:123 int-swim: mfizerlal o o - -’ g; 9 APPENDIX G METHOD OF SELECTION OF INTERVIEW MATERIAL Selection of the interviews used as input material for this study proceeded in the following manner: Five successive therapy sessions were tape recorded. Two of these sessions preceeded the day on which the psychological tests were administered, one was on the day the 0 was tested (the tests were administered two hours after this interview), and two of the recored interviews followed the testing. In selecting the interviews it was necessary that they be clear recordings; this requirement eliminated the interview conducted on the day of testing--the recording was of poor quality. While listening to the remaining four recordings, it became apparent that the introduction of the tape recorder had inhibited 0's spontaneity during the first recorded interview; P . ”L? made mention of this in the second recorded session. For this reason the first recorded interview was eliminated. It had also been decided by the author that the two interviews selected should be as close to the test date as possible. This requirement was an attempt to have 0 "similar" on these three occassions. Further, it was -100- felt that one of the selected interviews should have pre- ceeded the testing and the other followéd it. Accordingly, the second recorded interview--Session 11--and the fourth recorded interviewa-Session 13--were selected. These interviews became I' and I", respectively, Following the selection of the interviews, both ses- sions were re-recorded in their entirety with the ther- apist's remarks erased. -101- APPENDIX H RATIONALE FOR EXCLUSION OF ATTRIBUTIVE SIMILARITY SCORE In the early studies done on empathy--e.g., Dymond's study in 19H9 (2#)--in which a "prediction methodology" was used, the accuracy of the predictor was taken as a measure of his empathic ability. Hastorf & Bender (7,53) reported that this measure, which they termed "raw empathy," was significantly related to the similarity existing between J and 0. They suggest as a remedial measure, subtracting AS from ACC; the resulting score--"refined empathy"--would not be confounded by‘projection".1 Gage & Cronbach (#5) in an extensive review of the methodological problems involved in the computation of "refined empathy" scores concluded: "Clearly Bender & Hastorf did not arrive at a measure of accuracy independent of AS and RS." Hathaway (5%), moreover, questions the applicability of the term "projection" as used by Bender & Hastorf: [They]...have argued that likeness between a clinician's self-ratings and his prediction of the client's self-ratings can be used as a measure of projection. This arguement is not rigorous, since this kind of predictor success is not necessarily derived by inter- 1 Hastorf & Bender (53) report a significant ositive cor— relation between projection and raw empathy, and a significant negative correlation between projection and refined empathy. ~102— personal mechanisms. Terms for defense mechanisms all relate to affective involvement, and a number rep- resenting interpersonal similarity is not unequivocal evidence for "projection." If the word projection is to have useful meaning, it signifies a motivated inference; the in- ference is derived from repressed material. If a participant checks a personality item about himself identically with his prediction of a target person, projection could be in- volved, but it is not a parsimonious explan- ation. Even empathy, which is less clearly identified assumes a "feeling into." Again, mere identity does not establish the psycho- logical state. In view of these considerations the ACC score was used as the measure of achievement. AS and RS were seen as pos- sible factors capable of influencing the magnitude of the ACC score. This formulation is in accord with the thinking of Baker & Sarbin (h), who maintain that the ACC score is "prob- abilistically multiple-determined." However, these authors raise a crucial question regarding the "correction" of ACC for the AS factor. Baker & Sarbin ask: What is the rationale which Justifies cor- recting for "projection" but not for con— ceptual rigidity, personal distance, para- taxic distortion, ethnocentrism, or any other variable that might possibly mediate achievement? ...To assert that a general tendency to attribute one's own attitudes to others, for instance, should not be allowed to affect accuracy scores because it does not reflect "real understanding" is to remove social perception from the realm of achieve- ment and make of it something mysterious and detached from behavior. Consequently, ACC scores were not corrected for AS. RS, on the other hand, enters into the rationale under- lying Hypothesis 2 (cf. p.21); therefore, it was retained. -103- 7‘ APPENDIXII EFFECT OF DIFFERENT LEVELS OF EXPERIENCE 0N ACC SCORES The following tables present the results of an analysis of the ACC data (based on cluster"scores") taking into account the different levels of therapeutic and diagnostic experience of the Js. Two seperate analyses of the data were carried out: in the first analysis the Judges' levels of therapeutic experience were considered; in the second analysis their diagnostic exerience was the focal point. A rho correlation between therapeutic experience and diagnostic experience yields a value of r = .699 which,for N a 32, is significantly different from zero beyond the .01 level (102). Despite his high correlation, when the eight Judges in each of the four experimental conditions were seperated into a high and low group (the median was used as the cut-off point in each of the four conditions for both variables) three of the four conditions had different Js in these groups when this dichotomy was made on the basis of the two variables. Only the I"D condition had the same four Js in the high andeOW'groups regardless of the variable used for dichotomization--therapeutic or diagnostic experience. In View of this finding an analysis was carried on both variables. In the following tables H and L will mean High and Low group for the "experience" variable under consideration. 40M- TABLE XT ANALYSIS OF I"D CONDITION JUDGES DICHOTOMIZED ON BASIS OF BOTH DIAGNOSTIC AND THERAPEUTIC EXPERIENCE* ' SUM OF . MEAN SOURCE OF VARIATION SQUARES .92 SQUARE ‘E BetWeen H and L .1786 1 7.1786 9.1775' Between Js in same Group (H or L) 1,1119 ‘6 Total Between Js 1.2905 ' 7 Between Sessions .1175 1 .1175 1.23%2 Interaction: Ses- sions X Experi- .0063‘ l .0063 --- ence Interaction: Pooled _ Js X Sessions .5212 go, .0952 Total Within Js '''' ..6255 '.;§ TOTAL 1.98oo 15 * Dichotomization on the basis of therapeutic and diagnostic eXperience led to identical placement of Js; i.e, same four Js appeared in the high and low groups regardles of variable utilized for dichotomization. The F of 9.1775 obtained for the High and Low groups, for 1 and 6 degrees of freedom,is significant at the .05 level. An examination of the mean ACC scores for these two groups reveals that the low experience group had the higher mean Agc'score. The E of 1.23%2 obtained for between sessions, for l and 6 degrees of freedom is not significant. 4059 TABLE CXI ANALYSIS OF D I” CONDITION JUDGES DICHOTOMIZED ON BASIS OF THERAPEUTIC EXPERIENCE ‘—‘—7 L, T f WI 'SU'MOF' MEAN SOURCE-OF VARIATION SQUARES g; SQUARE E Between Hand L - - .1922- 'l ' .1922 1.1755"- Between Js in same . Group _ o 2810 _6_ .163 5 Total Between Js .' 1.1732 7 Between Sessions .1562 1 .1562 2.0206 Interaction: Ses~ sions x Experi- .0032 l .0032 --- ence Interaction: Pooled Js x Sessions .h640 ‘6 .0773 Total Within J's ' ' ' .623A ‘_§_ TOTAL 1.7966 15 ‘ivNeither the E of 1.1755 for the High and Low nor the ‘E of 2.0206 for between sessions, for 1 and 6 degrees of freedom, reach the .05 level of significance. The results for this same group of Judges dichotomized on the basis of their diagnostic experience are presented in Table Xlon the next page. The E of 1.308% obtained for between sessions, for 1 and 6 degrees of freedom, is not significant. "106'“ _.___._ .___.._. - I g A .. a n “ .- 0 6 I 9 t l r n. a. TABLE XII ANALYSIS OF D I" CONDITION JUDGES DICHOTOMIZED ON BASIS OF DIAGNOSTIC EXPERIENCE ' i ' ' ' ’ SUM OF ' ‘ MEAN SOURCE OF VARIATION SQUARES g; SQUARE 3 Between H and L .0037 1 .0037 --- Between Js in same Group 1.2868 9 .21h# Total Between JS 7' 1.2905 A 7 Between Sessions .1175 1 .1175 1.308% Interaction: Ses- sions x Experience .0387 l .0387 _-_ Interaction: Pooled Js x Sessions .5323 6 .0898 Total Within JS 8 ' .66 ' § TOTAL 1.9866 15 407. TABLE XIII ANALYSIS OF I'I" CONDITION JUDGES DICHOTOMIZED 0N BASIS OF THERAPEUTIC EXPERIEN SUM OF MEAN SOURCE OF VARIATION SQUARES g3, SQUARE ‘2 Between H and L .010#~ 1 .OIOE --- Between Js in same Group~ .8832 '§- .1h73 Total Between JS ‘8' .89u3 ' 7 Between Sessions .1369 1 .1369 1.5211 Interaction: Ses- sions x Experience .0888 1 .0888 --- Interaction: Pooled JS x Sessions ,529% ‘6 .0900 Total'Within Js "" .2661 ‘,_§ TOTAL 1.6604 16' The‘E of 1.5211 for between sessions, for 1 and 6 degrees of freedom, is not significant. Table XIV presents the results of an analysis of the I'I" condition dichotomized on the baSis of diagnostic experience. This table is given on the following page. The is in Table XIV'of 2.2740 for between sessions and of #.#H51 for the interaction of sessions x experience, for l and 6 degrees of freedom, do not reach the required level of significance. ~108- TABLE XIV ANALYSIS OF I'I" CONDITION JUDGES DICHOTOMIZED ON BASIS OF DIAGNOSTIC EXPERIENCE SUM 0R ' MEAN SOURCE OF VARIATION v SQUARES g; SQUARE 3 Between H and L .0922 I 1 .0922 -—- Between Js in same Group 28021 6 .1336 Total Between Js V I .89H3‘ 7 Between Sessions .1369 1 .1369 2.27H0 Interaction: Ses- sions x Experience .2676 1 .2676 h.#%51 Interaction: Pooled JS x Sessions .3616 6 .0602 Total Within JS .2661 ' _33 TOTAL 1.6601; 15 ~109~ TAM IF" ANALYSIS OF 1"!“ 00mm: JUDGES memm 01! M13 0? W0 mm uh ,1 “In: an»: OF ”I 80mm or manner mass 4: seem Between 1! am! I. .0038 1 .0033 - W .18 in same 0m an: .6. "MO Total m Js . .2503 7 Between Sessions .0283 1 .0283 “- Intereatiem Sessions a: maritime .0000 1 .0000 - Intemtiau Fooled J: 1 Sessions M g .1172 Total mm 3: f .1312 .52 MAI: .9822 15 m None of the Es in Table .. XVI are significant. Table Xfiwmmmted on the 1’01le page, gives the restate of a second analysis on the 1"!" grain). For the data in this table the group we dieimtonised on the basis of diamwstic experimoe. The obtained g in Tainan? oi“ 3.30% so: the interaction leseions x WHOM, for 1 ma 6 degrees of freedom, is not significant. Tm TABLE XVI ANALYSIS OF I"I" CONDITION ‘ JUDGES DICHOTOMIZED ON BASIS OF DIAGNOSTIC EXPERIENCE A? ...,I . . . SUM OF. R .. .. . .. SOURCE OF VARIANCE SQUARES Q: SQUARE E Between ‘H and L ' ' .0001 ' ‘ '1 .0001 4-7- Between Js in same Group 0 2 §02 _6_ o Oh‘l? Total Between Js .2503 ' 7 Between Sessions .0283 1 .0283 --— Interaction: Ses- sions x Experience .2h98 l .2H98 3.3042 Interaction} Pooled Js x Sessions .k538 g .0756 Total Within Js .2312 ‘ _§ TOTAL .9822 19 The tables on the next two pages, Table XVIand Table Xqucontain the results for all the experimental conditions considered simultaneously. Table XVlgives the results when all the Js in all the conditions are dichotomized into high and low therapy—experience groups. For Table XVIIthe Js were similarly dichotomized on the basis of diagnostic experience. ~111- TABLE XVII ANALYSIS OF ALL CONDITIONS JUDGES DICHOTOHIZED ON BASIS OF THERAPEUTIC EXPERIENCE ' ‘ SUM OF I~IEAN SOURCE OF VARIANCE SQUARES. 'Qi SQUARE '3 Between COnditions .2209 3 .0736 ~-- Between H and L .1683 l .1683 1.2531 r_ 3 Interaction: Con- 3 ditions x Exper- .2167 3 .0722 --- 5 ience 3 Between Js in same ’ j . . 1‘ Total Between Js 3.8292 31 L: Between Sessions .5685 1 .5685 6.53”” Interaction: Ses- sions x Experience .0188 l .0188 --- Interaction: Ses- sions x Conditions .0830 3 .0195 -_- Interaction: Ses- sions x.Exper- .0585 3 .0195 --_ ience x Conditions Interaction: Podled 2 0880 22 .0870 Js x Sessions Total Within Js 2.8168 32 TOTAL _ 6.6A60 63 The,E of 1.2531 for H and L, for l and 2# degrees of freedom is not: significant. The‘F of 6.53n4 for between sessions, for 1 and 2% degrees of freedom is Significant between the .05 and .01 levels. ~112- TABLE XVIII ANALYSIS OF ALL CONDITIONS JUDGES DICHOTOMIZED ON BASIS OF DIABNOSTTC EXPERIENCE ._ . _ »w: .: "A'SUM 0F . ., ._ :‘MEAN . SOURCE OF VARIANCE SQUARES Q; SQUARE E Between Conditions ' .2209 3 ' ,0736 -_- Between H and L .152h 1 .152H 1.0971 Interaction:'Con- ditions x Exper— .1212 3 .OH0# -—- ience Between Js in same Group 3.33%2 22 .1389 Total Between Js. ‘ W ' 3.8292 ' 31 Between Sessions .5685 1 .5685 8.3357 Interaction: Ses- sions x Experience .3006 l .3006 4.%O76 Interaction: Ses- sions x Conditions .0809 3 .0269 -—- Interaction: ses— sions x Exper- ience x Conditions .2262 3 .075% 1.1055 Interaction: Pooled Js x Sessions 1.6MO6 22 .0682 Total Within Js """ 2.8168 ' 32 TOTAL 6.6%0 63 Of the tabled E values listed above "between sessions" is Significant beyond the .01 level; the interaction sessions x experience is significant at the .05 level, and the remaining two are not significant. An investigation of the significance of the interaction reveals the significance to bethe result of the "between sessions" variable. ~IT3~ ecu- APPENDIXIJ ANALYSIS OF DATA: SCORES BASED ON ITEMS The data reported in the body of this study are based on the scores attained by the Js when the clusters of items were used as the basis for the necessary computations. The data were also analyzed with the individual items used as the basis for the calculations. A rho correlation, based on the ranked mean ACC scores(for each Judge)attained byfi these two methods of analysis, was computed. The obtained value, r: .830 is Significantly different from zero at the .01 level. Thus there would appear to be a significant degree of similarity between the way the Judges (basedmmean"ACC)are ranked by these two methods of analySis--clusters Kg items. Table XIXL (p.105) gives the correlation scores for all the Judges,when computed on the individual items. A comparison of this table with Table IVI (p.h9) reveals the differences in the obtained values of the scores resulting from the two methods of computation. When based on clusters the ACC scores range from -.217 to .733; when based on items the ACC scores range from -.102 to .hh9. The cluster correlations are based on 17 "need scores", the item correlations are based on 70 "item scores". Twenty of the sixty-four ACC scoresreach the .05 level of Significance or beyond,in the former method; 30 do in the latter one; .H68 and .23h are the respective .95 levels. ...lzlh'. TABLE XIX; ACC AND RS CORRELATIONS FOR ALL JUDGES (Based on Items) H ‘1 SESSION GR. , GROUP JUDGE ONE ACC TWO MEAN RS A ".033 0010 -0009 B .375** .076 .300* c .3#6** .352s* .360** D .323** .2lh .105 I"D E -.027 -.056 .103" F .237* .29A* .395** o .2: * .386**, .289* H .9 8 -0061 0220 Mean -- .200* —-- .15h .179 I -.O33 .205 . 17** J 0116 001 O 26** x .228 .20 .2%3* L .128 .231 .h35** D I" M .180 .h15** .111 N .369** .306** .269* O .180 .20 . 86** P .h21** .386** . 38** Mean -- .20 * --- .2h9** -.226* R .2 * cl . .326** s .363** .266* .271* T‘ .3 2** .228 .231 U .2 3* .02 .306** 1'1" v .223* .328 .228 W' 02 I o 1 ** “0130 x .395** .%49** ~21E~ Y .0# .215 .323** Mean -- .27 **--- .228* .251** AA .3#6** -.102 .312** BB .3A6** .217 .360** CO .203 .323** .346** DD .002 .111 .300* I"I" EE .27A# .225* .363** FF 0251+,“ “0093 02 7* GG .312** .312** .1 5 HH ' .1h2 .O9M- .237* Mean -- .237* --~ .139 .188* TOTAL MEANS .211* .229* .l9n* *: Significantly different from zero at .05 level ** Significantly different from zero at .01 level -115. IF" Despite the significant correlation found between mean ACC scores as computed by the cluster and item methods, when an analysis of variance is applied to the data obtained by the latter method (as shown in Table AXX' ) no significant results are found. Presumably; this is the consequence of the lower reliability of the ACC scores when they are based on correlation of the individual items, cf. Table VII,p. 5A. The discrepancy in the results obtained by the use of these two methods of computation, suggests that the scoring method,used in studies employing Q-technique, is a critical factor in evaluating the results obtained. TABLE XXL" ANALYSIS OF‘VARIANCE OF ACC SCORES* FOR FOUR GROUPS OF JUDGES UNDER DIFFERENT EXPERIMENTAL CONDITIONS WITH TWO SESSIONS FOR EACH GROUP J. —_— - #- SUM OF ’ ' MEAN SOURCE OF VARIATION SQUARES Q; SQUARE ‘2 Between Conditions .0600 3 .0200 .5780 Between Js in same 3 Group 0 220 [‘2‘ _§ 0 0311'6 ‘ Total Between Js 1.0305 1 31 Between Sessions .0213 , 1 .0213 1.5106 Interaction: Ses- sions x Conditions .0275 3 .0158 1.1205 Interaction: Pooled Js x Sessions .3252 .2§ .0121 Total Within Js ' .2622 " 32 TOTAL 1.A9#7 63 * COrrelations based on individual items -116- I The Es of .5780 and 1.1205 for "between conditions" and the interaction sessions x conditions, for 3 and 28 degrees of freedom, are not Significant. The "between sessions" 2 value of 1.5106, for l and 28 degrees of freedom, is not significant.' Thus none of the results obtain,when the data are based on correlations between the individual items. The statistical significance is found when correlations based on clusters of itemS-—needs--are utilized. ***** *** -117- R 0 0M USE ON LY if": TN 9" .'\ L. ,5