University of Southern California Dissertations and Theses

Intellect After Lobotomy In Schizophrenia: A Factor-Analytic Study

(USC Thesis Other)

Intellect After Lobotomy In Schizophrenia: A Factor-Analytic Study

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INTELLECT AFTER LOBOTOMY IN SCHIZOPHRENIA A FACTOR-ANALYTIC STUDY by Richard de Mille A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Psychology) June 1961 UNIVERSITY O F S O U T H E R N CALIFORNIA GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES 7. CALIFORNIA T h is dissertation, w ritten by Richard de Mille under the direction of /i.ls.Dissertation C o m mittee, and a p p ro ve d by all its m em bers, has been presented to and accepted by the D ean oj the G raduate S ch o o l, m partial ) ulfillm ent oj requirements ja r the deejree of D O C T O R O F P H I L O S O P H Y Dam Dale. r J p / C J .... DIS^ERTATION 0, 0 MMITTEE i rman ACKNOWLEDGMENTS The author wishes to express his indebtedness and gratitude to the members of the dissertation committee, and Dre P. R. Merrifield, for meticulous guidance and generous consultation throughout the project. Sincere thanks are due the Managers, Directors, and staffs of the Veterans Administration Center, Los Angeles, Brentwood V. A. Hospital, Palo Alto V. A. Hospital, and Sepulveda V. A. Hospital; in particular to Drs. F. Harold Giedt, Thomas W. Kenneily, Joseph Rubinstein, and John R. Schlosser; and to Dr. Frank J. Kirkner of Long Beach V. A. Hospital. Their cooperation and assistance in the job of finding, selecting, and testing subjects were indispensable. Special thanks are due Dr. Harry M. Grayson of Brentwood V, A. Hospital, and Dr, Paul McReynolds of Palo Alto V. A. Hospital, who made possible and greatly facilitated the two major testing programs in the project. The author is grateful to Dr. Leon Epstein, of the California Department of Mental Hygiene, for his aid and encouragement at the outset of the study; and to the Superintendents and staffs of Agnews, Atascadero, Camarillo, De Witt, Mendocino, Metropolitan, ii Modesto, Napa, Patton, and Stockton State Hospitals for their cooperation, assistance, and hospitality during the data-gathering’ period. In particular, thanks are due to Drs. Trent E. Bessent, William A. Cook, Jr., Sarah Counts, Hershel Fogelson, Bernard Meer, Margaret Noszlopi, Gordon L. Riley, Benjamin Siegel, Wilson M. Van Dusen, and Seymour L. Zelen; to Gerald A. Green; and to Drs. Kristian Johnsen and Frances Sheridan. The author feels fortunate in having had the opportunity to consult, with Dr. Walter Freeman in the early stages of data collection. Subjects not tested by the investigator were tested by Alice Beach, Stephen Carraway, Harry Carritte, Beverly Collins, Thomas Coke, Raymond Conatser, Chester Cooley, Sheila Farley, Joseph Fischer, Edward Gould, Jacqueline Griffin, Barbara Griswold, Millard Madsen, Ira Nathanson, Alexander Quenk, Rollin Rose, H. William Safford, Frederick Stoller, Ruby Thompson, and Dr. Margaret Noszlopi. The Project for Studies of Aptitudes of High- Level Personnel, University of Southern California, gave permission for the reproduction and use of a number of its experimental tests, and made its graphic equipment available. Extensive use of calculating equipment was provided by the Psychology Department of the University of Southern California and by the Monroe Calculator Company. Correlations and analytical factor rotations were computed by the Western Data Processing Center, School of Business, University of California at Los Angeles, whose aid is gratefully acknowledged. This investigation was carried out during the tenure of a Predoctoral Fellowship from the National Institute of Mental Health, United States Public Health Service. iv TABLE OF CONTENTS ACKNOWLEDGMENTS LIST OF TABLES Chapter I. INTRODUCTION ......................... . II. BACKGROUND: INTELLECTUAL CONSEQUENCES OF BRAIN LESIONS AND ABLATIONS . . . Human Studies Animal Studies III. BACKGROUND: FACTOR-ANALYTIC STUDIES . IV. THE PROBLEM . ........... V. PROCEDURES.......... ................. Description of the Tests The Subjects The Pilot Project Administration and Scoring of Tests Treatment of Data Statistics VI. THE FACTOR ANALYSES ................. Extraction of Factors Rotation of Axes Interpretation of the Factors VII. INTERGROUP COMPARISONS ........ Hypothesis Testing Factor-Defined versus Task-Defined Tests VIII. INCIDENTAL FINDINGS AND FURTHER USE OF DATA . ........................ Chapter Page IX. DISCUSSION............................ 107 X, SUMMARY.............................. Ill BIBLIOGRAPHY .................................... 116 vi. LIST OF TABLES Table Page 1. Group Matching on Continuous Variables . 59 2. Subgroup Matching on Continuous Variables............................. 60 3. Categorical Group and Subgroup Matching 62 4. Means, Standard Deviations, and Reliabilities of Scores, Experimental G r o u p ............... . 72 5. Means, Standard Deviations, and Reliabilities of Scores, Control Group 73 6. Correlation Matrix, Experimental Group . 75 7„ Correlation Matrix, Control Group . . . 77 8. Centroid Matrix, Experimental Group . . 82 9. Centroid Matrix, Control Group ..... 83 10. Rotated Factor Matrix, Experimental Group .................... . 86 II„ Rotated Factor Matrix, Control Group . . 87 12, Comparison of Means and Variabilities of Main Battery Variables ...... 96 13o Comparison of Means and Variabilities of IVechsler-Bellevue Scale, Form I, Subtests and IQ's .......... 98 14. Comparison of Means and Variabilities of Factor-Analytic Tests ........... 99 vii CHAPTER I INTRODUCTION A study of intellectual consequences of pre- frontal lobotomy is readily classifiable as a study in physiological psychology. The province of physio logical psychology is broad, however, and a single study must have a very limited scope. The first task of this introduction is to show where the present investigation fits in. Among the realms of variables comprising physiological psychology are those relating to bodily function and behavior. As a convenience, an experi menter may limit these two categories to independent variables representing induced variations in bodily function and dependent variables representing observed variations in behavior. Variations in bodily function to take one side of this relationship, may be induced through modifications that fall into two broad classes immediate physiological modifications, and anatomical modifications. Physiological modifications, such as those produced by chemical agents, electronic applica tions , and other means whose effects are usually reversible, do not figure in the present experimental design and will not be discussed in any detail, 1 2 though it is not implied that they lack importance; on the contrary, it is felt that future studies will find some of their most rewarding experimental variables in that category. For the present, irreversible ana tomical modifications take the central role. The kind of anatomical modification used in a physiologico-behavioral study depends partly on whether the subjects are animal or human. The majority of animal studies make use of surgical preparations, whose sole purpose is to provide experimental variables. Investigators using human subjects have to be content with modifications that are experimentally less satis factory— less exactly specified, less accurately carried out, less frequently verified, and more often subject to the intrusion of unwanted influences. Even when such modifications serve as experimental manipulations, they have primarily a therapeutic and only secondarily a scientific purpose. On other occasions, they may only be observed, without possibility of experimental manipulation and with, at best, difficult problems in statistical treatment. Of the many possible kinds of bodily modification, only brain modifications will be discussed below, since the present problem involves lobotomy. 3 Among conditions that may be observed natural- istically and controlled statistically are exogenous and endogenous feeble-mindedness in children, organic psychoses, degenerative diseases, and cases of trauma. In a few studies, post-mortem examination may provide verification of the diagnosis, but most go unverified. There are situations in which surgical inter vention may provide data. Removal of tumors, repair work consequent to traumas, and ablation of pathological tissue are examples. When sufficient numbers of cases are observed, something approaching the discreteness of experimental manipulation may be achieved through statistical means. Serious problems in defining and selecting the sample, however, limit the usefulness of this kind of data. The advent of psychosurgery (16, 18) in which lesions of the brain are made in order to relieve the patient of intolerable pain or mental anguish, or to render him behaviorally more tranquil, brought about a somewhat more favorable experimental situation0 In psychosurgery patients are subjected to a uniform operation, or to a limited number of uniform opera tions. Three main classes of psychosurgical operation are: destruction of subcortical structures, destruction of cortical structures, and disconnection of cortical 4 from subcortical structures. It is in the third of these classes that the brain modification used in the present study belongs. Lobotomy has been chosen here neither for theo retical nor for experimental reasons, but for statis tical, practical, and historical reasons. The empirical background for theoretical statements about intellectual effects of lobotomy is equivocal. A purpose of this study is to probe for evidence that could reduce that equivocality. Experimentally, the operation is neither well specified, reliably performed, open to verifica tion, nor free from misleading sequelae. Statistically, however, lobotomies present the advantage of having been performed in large numbers during a relatively short interval for relatively unvarying clinical reasonso Practically, the patients upon whom they were performed are still available for psychological testing. Lobotomy, moreover, is a landmark now passing into medical history 5 the possibility of using loboto- mized patients in large numbers is one that will diminish rapidly. The past two decades have seen many studies of lobotomized patients giving rise to contra dictory conclusions. It would be a good thing to produce some resolution of the controversy, on its home ground, before it is too late. For these reasons 5 lobotomy has been chosen as the experimental manipula tion, in spite of its shortcomings as an experimental procedure. The primary purpose of the study is to demon strate refinements in the application of psychological tests to the measurement of behavioral correlates of physiological variables. The project will be amply successful if it encourages and facilitates programs of research in physiological psychology, both animal and human, and in abnormal and clinical psychology, that will take advantage of important developments in psychological mensuration— notably the developments of factor analysis and of networks of test instruments whose meaning has been clarified through the factor- analytic reduction of systematic, theoretically-oriented successions of large correlation matrices. It will be argued below that advances have been made in physiology and psychology that are quite comparable in their methodological significance and that call for reap praisal of the kinds of experimental design that may be needed for progress in physiological psychology. Further delimitation of the study must precede that argument, for only the physiological side has been dealt with, and attention must now turn to the behav ioral side, 6 All animal and human behavior is relevant to physiological psychology. Consequently, all psycho logical systems of classification and conceptualization of behavior stand ready to contribute to physiologico- behavioral studies. A single research program, much more a single investigation, must choose a restricted, internally consistent set of behavioral concepts and experimental operations if the results are to be interpreted with clarity. The experimental operations defining the dependent variables in this study are drawn from contemporary factor-analytic work with psychological tests of human aptitudes. Such specifi cation limits the scope of the variables a great deal, relative to the vast universe of animal and human behavior. Guilford8s classification of human behavior (27, 28), which is the contributing one here, lists needs, interests, attitudes, and temperament, in addition to aptitudes, as variables of behavior. The aptitude category is divided into perceptual, psycho- motor , and intellectual dimensions. The intellectual dimension, in turn, has five major divisions (28, 38): cognition, memory, divergent production, convergent production, and evaluation, four of which are repre sented in the present study. Superimposed on these 7 divisions are the aspects of intellectual content and product, creating a three-dimensional matrix type model, which is capable of generating predictions about 120 or more factors of intellect, of which 55 have already been empirically defined. Eight known factors of intellect were chosen as reference points for the interpretation of results in this study and, before that, to indicate an appropriate battery of tests. The dependent variables are 16 factor-defined test scores. Most previous studies of psychological effects of brain modification have depended upon clinical observations, "general intelligence" tests, or factorially undefined specialized tests. In the main, the results have been disappointing, as a survey of the literature will show. CHAPTER II BACKGROUND: INTELLECTUAL CONSEQUENCES OF BRAIN LESIONS AND ABLATIONS Human Studies Two general summaries of the problem of brain lesions and concomitant personality change reveal little progress during a twelve-year period. White (104), in 1956, quotes a number of Cobb's 1944 specu lations as still contemporary, while Cobb (7) refers to the same major aspects of the problem that face us today. The contributions of animal studies to psychosurgery, the difficulties in producing the intended lesions in the brain and in controlling side effects, the question of the most useful kinds of psychological measurement, the localization contro versy , and the possible contribution of advanced correlation techniques— all are explicitly stated by Cobb, who mentions Lashley, Hebb, and Thurstone in contexts quite suited to the present study. Progress during the past twenty years, in fact, has been largely invisible, because the common impli cations of advances in discrete areas do not auto matically announce themselves. They must be noticed, thought through, stated, and published, if they are to be generally recognized. Unnoticed implications and unpublished statements have obscured a growth that should now be brought into the light. A large literature backs up this contention. Fortunately, much of it has been well reviewed. A number of references below are to surveys that cover hundreds of relevant articles. The most extensive published review of studies of psychological consequences of brain lesions and ablations is found in two articles of Klebanoff = > The earlier of these (54) begins with Fritsch and Hitzig and carries the search forward to 1941. One section considers qualitative studies of traumas, tumors, and ablations, occurring in the various cerebral lobes; another takes up quantitative studies and the applica tion of psychological tests, A distinction is made between "unspecialized" tests such as the Stanford™ Binet and "specialized" tests such as tests of Goldstein, VVeigl, and Bender. The former are said rarely to demonstrate psychological change due to brain damage, while some of the latter are said to be moderately successful in demonstrating change. It is suggested that an understanding of the nature of psychological consequences of brain modifications, and the determination of their physiological bases, 10 await the development of new special tests that will be capable of analyzing global intelligence and other general behavior modalities into their component parts. The later article (55) carries the review forward to 1952 and is organized around etiologies and symptoms rather than around the major cerebral lobes. The authors report small progress during a ten-year period with the issue of general intellectual effects of brain injuries or diseases. Most of the research was confined to the VVechsler-Bellevue Intelligence Scale; attempts were made to tease out patterns of subtest scores that would be indicative0 Many investigators reported that the Digit Symbol, Digit Span, Block Design, and Arith metic subtests were especially sensitive to brain injury. Diagnostic implications were clouded by the supposed sensitivity of these same subtests to anxiety and emotional disturbance. Klebanoff et al. criticize much of this research as excessively test bound and empirical and recommend that a theoretical approach be stressed, as in new departures by Cattell, Hebb, and Halstead. One weakness of "general intelligence" as an indicator is pointed out (55) in a reference to critical case studies that show no decrement or even show an increment in intelligence after large excisions of 11 cortical tissue. This paradox is partially explained by Hebb (46), who states that behavior is more dis rupted by the presence of pathological tissue than by the absence of tissue. A clean excision of pathological tissue may leave the patient more able than he was just prior to operation. Nevertheless, large cortical ablations ought to have some noticeable effect on intellect, and if they appear not to, the reason may be sought in the nature of the testing instrument. Klebanoff et ai, call the literature on psycho surgery disconcerting. They attribute inconsistent results to a number of defects of design, such as lack of preoperative measures, questionable reliability of preoperative measures, uncontrolled intrusion of extraneous influences like attitudes of personnel and relatives, an inadequate sample of subjects, absence of a control group, unstandardized procedures, dis regard of practice effects, and absence of statistical evaluation. An additional hazard, particularly impor tant when samples are small, is the unreliability of the anatomical modifications produced by lobotomy, topectomy, or thalamotomy; reports by Meyer and MeLardy and by Mettler (67, 68) are cited. A revealing compilation of psyehosurgical studies is presented (55) s including’ nine lobotomy 12 studies, one topectomy study (Columbia-Greystone), and one mixed study. Pre- and post-tests were made with the Wechsler-Bellevue or the Stanford-Binet. Subjects were psychotics, neurotics, or normals with intractable pain. Klebanoff et al. conclude that the evidence strongly suggests intellectual impairment following frontal-lobe tissue destruction or dis connection. Apparent increases in IQ are attributed to practice or other extraneous effects. Data from psychotic subjects are held to be relatively invalid; all four of the statistically significant results, which are in the direction of decrement, are from neurotic and normal subjects. A single study of five normals showed an IQ decrement of 20 points. Another compilation is made of studies using the Porteus mazes in the same way. In spite of the susceptibility of the test to practice effects, five out of seven studies show an apparent decrement, which is significant in two studies. Later post-tests showed a restoration of performance level, again attributed by Klebanoff et al. to practice. In their general comments, Klebanoff et al. mention a decreased emphasis on problems of localiza tion, due partly to failing optimism about the utility of psychological tests in such problems. The impasse 13 is seen as a result, on the one hand, of inadequate specification of the locus and sequelae of the injury and, on the other hand, of insufficiently specific psychological tests. Two lines of test development are suggested: in laboratory test techniques, such as critical flicker fusion and tachistoscopic presenta tion ; and in newly defined relations between tests and the functions they are supposed to measure, as in the factor-analytic approach, first applied by Halstead. These comments reflect the same general con ception of the problem that is being developed in the present study. The emphasis given by Klebanoff et al. is different, however, for they see the factor-analytic approach as one challenge among many, while the present study takes the position that it is unlikely that any behavioral indicator not factorially defined will reflect any unitary function with sufficient clarity to be useful in demonstrating correspondence between specific brain and behavioral modifications„ The factor-analytic approach is held to be a necessary condition for a successful attack on physiologico- behavioral problems when subjects are human and behaviors complex, and may also be necessary in many problems involving animals or supposedly simple behaviors. Klebanoff et al. do not single out the 14 test-bound or nonfactorial approach as a major con tributor to paradoxical results in the literature. The present study takes the position that the chief fault, on the behavioral side, may have been the use of undifferentiated indicators to measure relatively unitary functions, and, further, that the factor- analytic approach or some equivalent is indispensable to the definition of indicators that are appropriately univocal. In a 1953 study, some of whose subjects are subjects in the present investigation, the Wechsler- Bellevue Intelligence Scale was applied "to determine whether components of the IQ shift following lobotomy, even in those cases where the IQ remains relatively constant" (63:229). Since the sample was small and psychotic, and since practice effects were not con trolled, it is not surprising that no significant differences were found. The purpose of the study deserves recognition, however; for it is an effort to demonstrate a change in the structure of intellect when mean scores on complex indicators do not change. In one of the most significant papers bearing on the problem at hand, Hebb (46) reports the case ox K* ? * 1. , a complete, bilateral frontal lobectomy, and makes comparisons with other cases of tissue 15 removal. He finds that the case of K. M. alone meets the criteria for suitable demonstration of tissue removal, uncomplicated by residual malfunctioning tissue or by uncertain specification of the line of section. Subtle deficits seemed present in K. M.'s personality after lobectomy, but no impairment could be seen on psychological tests. Discussing prefrontal lobotomy, he says that, although operations on the frontal lobes are followed by defects in almost all cases, the defects are due to residual pathology and not to tissue removal or disconnection per se. "Pre frontal lobotomy is a landmark in psychiatric therapy, but has yet to provide any interpretable evidence concerning what goes on in the normal frontal lobes" (46:21). This evaluation is echoed fourteen years later by White (105), who says that one of the dis heartening features of lobotomy was its failure to make the expected contribution to scientific knowledge. Hebb takes to task those who report no deficit on psychological tests and who conclude therefore that the frontal lobes have no intellectual functions. "It must be assumed that the region has an important role in behavior. . . . Failure to prove that it has certain functions does not show that it has none" (46:23). He takes the position that we have no 16 information about frontal lobe function, partly because of failure to specify brain modification accurately, and partly because of inadequate psychological measure ment . The disadvantage of using psychotic subjects in lobotomy studies is underlined by Hebb (47), who points out that most of the intellectual deficit has already occurred, before lobotomy, as a result of the psychosis. He cites an investigation (84) in which premorbid measures of intelligence were obtained for a small number of lobotomized patients and were com pared with pre- and post-lobotomy measures. Tests given just before and after lobotomy showed nearly identical means, while the premorbid mean had been 18 IQ points higher, a statistically significant differenceo "There can be no doubt," Hebb concludes, "that lobotomy in a normal brain would induce serious defects of problem solving" (47:180). Related studies are cited by Rosvold (83). The present investigation uses psychotic sub jects and is therefore subject to the difficulty discussed by Hebb, Rosvold, and others. It was thought that this handicap could be overcome by using tests that were more univocal and by treating the data factor-analytically. 17 Thirty-six nonpsychotic subjects were studied by Tow (97), before and one year after frontal lobotomy. The subjects were chosen for interest, cooperativeness, and relative preservation of personality. There was no control groupo Contrary to expectation, nonintellectual indicators showed little or no change, while intellec tual indicators showed significant decrements. The Raven Progressive Matrices and the Porteus mazes were among the tests showing impairment. Probably the most celebrated investigation of the effects of psychosurgery was the first Columbia- Grey stone study (67), which gave birth to copious publications and stirred up an extended controversy. Though the experimental plan involved opening the skull and excising small portions of frontal cortex, in an operation called topectomy, and v/as not concerned with lobotomy, the research design and the conclusions of some of the investigators have implications for the present study. Deese and Morgan describe the study as follows: Forty-eight subjects were intensively studied by a team consisting of every type of scientist who presumably could contribute something. . . . Twenty-four patients were controls, and 24 experi- mentals were subjected to a variety of bilateral partial frontal ablations. . . . In the 24 operated subjects, the 11 Brodmann areas rostral to area 4 were systematically removed in various combinations. . o . The data are so summarized as to provide 18 information concerning anatomy, physiology, psychology, and psychiatry. On the specific psychological tests there were very few detectable changes. A few temporary and isolated changes were suspected, though these did not reveal any consistent or meaning ful pattern .... The findings in this study are almost uniformly negative. They have the function, however, of pointing out that there is no existing theory of the function of the frontal areas in human beings which seems to flourish in the presence of a good deal of relevant data. The authors of this study were able to demolish a consider able number of notions about the functions of the frontal lobes .... (12:202) Some similarities to the present investigation are of interest. Mettier and Curry (67) report difficulties in producing the desired incision, lead ing to uncertain specification of the anatomical modification; lobotomy is even more subject to this handicap. The subjects were comparable to those in the present study, in some respects: they were psychotic, they varied in age from twenty to sixty years, and showed an IQ range of 55 to 131. The present sample are schizophrenic, varying in age from thirty to sixty; the veterans show an IQ range of 67 to 128. An obvious difference is in the size of the sample. The Columbia-Greystone sample is very small, while the present sample may be called intermediate. 19 In the earlier study, group matching failed at several points: the groups were not well matched for age, diagnostic entity, or number. One-third of the sub jects were lost to psychological testing during the course of the study, but more than twice as many control as experimental subjects were lost. In such a small sample, the loss of even a few subjects may markedly increase the biased nature of the sample. Commenting on these difficulties, one investigator says, "We had to depend on the scientific art of arriving at sufficient conclusions from insufficient data" (67:177). Another difference was marked intragroup heterogeneity of diagnosis. Of 48 subjects, 36 were schizophrenic; the rest were manics, depressives, involutional3 or other. The research designs differ radically in their choice of anatomical modification,, In the present study, 150 patients had the same operation, performed by a limited number of techniques. In the Columbia- Greystone study, the number of different operations was equal to the number of operated patients. No two subjects had, or were intended to have, the same operation. Even when subjects are combined into categories, confounding the discrete effects of locus 20 with interactions among loci, conclusions about the different effects of the various removals are reached on the basis of groups of from three to six subjects. Unlike the present investigation, the Columbia- Grey stone study was a before-and-after study, with all the control difficulties of that design. Patients were selected by medical rather than experimental criteria. Extraneous differences introduced by the fact of having an operation had little time to dis appear; the same was true of nonoperative differences in treatment. Conditions favored practice effects, and efforts to control them were inadequate, as has been pointed out (55). One hundred psychological tests are considered in the Columbia-Greystone report. With a few possible exceptions they are factorially highly complex. The chief intelligence test was the Wechsler-Bellevue Intelligence Scale. Though one factor analysis is reported, no factor-defined instruments were included. In view of the small sample, the matching difficulty, the superabundance of surgical procedures, the heterogeneous diagnoses, the probable intrusion of extraneous influences, and the complex and undefined instruments, in addition to unavoidable anatomical inaccuracies and the use of psychotic subjects, it is 21 hard to agree with Deese and Morgan that "the authors were able to demolish a considerable number of notions about the functions of the frontal lobes" (12:202). One Columbia-Greystone investigator states that, faced with the Columbia-Greystone data, "no existing theory or hypothesis dealing with the psychologic significance of the human frontal lobes is tenable" (67:496); ". . . our critical experiments definitely prove that these [intellectual, categorizing, abstracting.1 functions are not primarily connected with frontal lobe tissue" (43:263). Data showing nothing positive and deriving from such epistemically tenuous procedures can hardly be held to have performed the difficult, and in some senses impossible, task of negative proof. What the Columbia-Greystone study did seem to render netenable, or to demolish, was the notion that research of such design was likely to enhance physiologico- behavioral theory. The second Columbia-Greystone study (55, 68) will not be considered here. In the third Columbia- Greystone project (59), a comparison is made of superior and orbital removals from the human frontal cortex. In reviewing the work, Garner says, "Few of the wide range of studies reported here show a dif ferentiation between the two varieties of operation; 22 indeed, a few show significant differentiations between operated and control patients" (19:404). In spite of an increase of numbers and other improvements of design, little difference between groups is noted. Character istically, the Porteus mazes are more sensitive to changes than most of the other tests. Conclusions are drawn by Sheer (19) from the three Columbia-Greystone projects, to the effect that with circumscribed lesions there are no permanent intellectual losses in psychotic patients, and that temporary losses are a function of temporary physio logical changes in the brain and not a function of any particular operation. The qualification that there are no permanent losses in psychotic patients is a welcome one, and yet a loss was demonstrated by Smith and Kinder (91), who retested the available survivors with the same battery of tests, eight years after operation. In contrast to the earlier findings, eight of the fourteen instruments differentiated the groups significantly. Subjects were dichotomized as operated and control, and as older and younger. Superior versus orbital operations constituted a third dichotomy. Superior removals had a markedly greater effect than, orbital removals. Age made little difference in the 23 effect of superior removals but worked to the dis advantage of the older subjects with orbital removals. The authors conclude (89, 90, 91) that four considerations related to impairment are: the specific site of the lesion, the age of the subject, the nature of the measure, and the postoperative interval. An explanation of the delayed appearance of the impairment, and of the different effects of superior and orbital removals, is sought in neurological studies showing correlation between amount of brain degeneration and time elapsed since psychosurgery. Teuber reviews literature that calls in question the preeminence of the frontal cortex in mediating higher mental processes and emphasizes the importance of posterior lesions. He concludes: The present evidence indicates that, on certain tasks, frontal as well as postcentral lesions can produce significant deficits in man. But the evidence does not mean that one could not identify specific symptoms of anterior or pos terior lesions if the tasks were appropriately constructed .... (93:285) He cites a study by Teuber and Mishkin in which differ ent tasks in perception and manipulation of the vertical were affected by anterior and posterior lesions. This success in demonstrating differential effects lends weight to the suggestion (55) that laboratory techniques have an important contribution to make in the search 24 for unitary behaviors. A purpose of the present study is to demonstrate the use of similar unities arrived at by mathematical means. Milner (69) reports that aphasias have been linked to left temporal lobe lesions by hypotheses (a) that aphasia is merely a manifestation of a more fundamental and general intellectual disorder, or (b) that there may be loss that does not affect all aspects of intelligence equally, so that subjects do poorly on some tests and wel^ on others, or (c) that aphasia is a mere loss of symbols, producing difficulties in communication but no intrinsic difficulty in problem solving. Milner reports that the second hypothesis has the most empirical support and concludes: . . , in aphasia the pattern of test performance is likely to be one of severe deficit on verbal tests and relatively mild deficit on nonverbal ones. . . . the degree of this disparity between verbal and nonverbal scores varies markedly from person to person, and with the tests used. These differences, which are very striking, are almost certainly related to the site of the lesion. . . . An additional factor may be the degree to which the individual normally relies on verbal cues. (69:55) Her conclusions may be paraphrased in factor-analytic terms. Aphasia is characterized by large score decrements on tests representing verbal factors and small score decrements on tests representing nonverbal factors. Individual differences in the disparity 25 between verbal and nonverbal seores may be attributed to three sources: (a) variation from individual to individual of the factorial composition of the given task, (b) variation from test to test of loadings in relevant verbal and nonverbal factors, and (c) varia tion from individual to individual of the degree of coincidence of the lesion with the exact structural unity that mediates the factor-defined ability in question. A factor-analytic approach seems called for in the area discussed by Milner. Scherer, Klett, and Winne sum up a five-year project in which small numbers of lobotomy and control patients were tested with forty measures of "function ing efficiency," culled from an original array of 106 measures. Inspection of the table (87) shows that, before operation, the patients to be lobotomized were apparently superior to the controls on twenty-three of the forty measures; five years after operation, they were apparently superior to the controls on thirty- three measures. "In a global sense," the authors say, "the lobotomized patients demonstrated an increase in functioning efficiency," though they point out that "there is little consistency [from one year to the nextJ in the particular measures on which it is manifested" (86:293). 26 These authors do not attribute the suggested superiority of the lobotomized group to an improvement in any test-represented ability. They do, however, discount undesirable effects of lobotomy, inferring that either the frontal lobes do not mediate the func tions studied or lobotomy does not interfere signifi cantly with the function of the frontal lobes. In a 1953 study of 25 lobotomy patients, some of whom are subjects in the present investigation, Ruja (85) applied the Wechsler-Bellevue Scale, Form I, one month before operation and six months after operation, and Form II one month after operation. There was no control group. No enduring change was found on any subtest or IQ score. Nineteen person ality indicators also showed no enduring change. Between 1951 and 1955, in another before-and- after study (65), a varying battery of tests was applied periodically to groups of from 23 to 56 lobotomy patients, some of whom are subjects in the present investigation. A memory test and the Digit Symbol subtest of the V/echsler-Bellevue Scale were included. There was no control group, and the data did not justify a statistical treatment. The authors noted an apparent improvement in retentive and attentive (Digit Symbol) capacities following 27 stabilization after lobotomy. Pribram makes a statement that expresses both the frustrations of those who have tried to analyze the subtle effects of psychosurgery and their reluctance to let go of the problem until a solution is achieved. He says, in part: There are those who hold that the frontal cortex of man is the "organ of civilization" and that tampering with this structure comes close to criminal action. On the other hand, there are those who claim that no consistent effects are ever observed to follow frontal lobectomy or leukotomy .... Only a few standard psychological tests have been successful in demonstrating any change in the psychosurgical patient. . . . Yet, psychiatrists and persons who are in close contact with a post- lobotomy patient have no difficulty in spotting the fact that some important change has taken place in the patient. [Alleviation of symptomsj is at a price, and the price is so hard to define. Perhaps the difficulty lies in the approach to the problem. (76:19) The available evidence, Pribram suggests, would justify an investigation of the frontal cortex in terms of intentions, the projection and carrying through of a complex series of actions. Evidence continues to mount confirming the necessity for a more analytical and multivariant approach to perceptual, intellectual, and motor con comitants of brain alteration. Rosvold (83) reviews twenty-two studies with human subjects, where conflict- 28 ing results are apparently due largely to the selection of behavioral indicators. An array of psychological tests is recommended by Burdock et al. (2). The tests vary from laboratory techniques to complex intellectual tasks, and have been found sensitive to differences among patient groups, including groups with brain pathology. It is a good thing to explore the meaning of such tests, but if the exploration is innocent of all analysis of test interrelationships, it must fall short of the desirable in test development. A task™ defined approach to testing must in the future be ancillary to the main attack on the unknown structure of behavior and the undiscovered relations between the behavioral and other realms. Animal Studies In the remainder of this section, some important points where animal studies bear upon the present investigation will be indicated. Illustrative articles and studies will be cited. Stellar (92) contrasts the greater control over locus and extent of lesions, and the simpler behavior, found in animal problems with the lack of control and complex behavior found in human problems. Manifestly, experimental differences are great, and yet animal 29 and human studies have indispensable contributions to make to each other. Teuber states: Work on animals has profited from assimilation of testing techniques originally devised for man and work on men with cerebral lesions has been equally aided by the application of tasks used earlier for nonverbalizing organisms. We are seeing the beginning of a truly comparative approach to the effects of cerebral lesions or stimulation. So far, results in man and sub human primates are in better agreement than we dared to expect. (93:287) Harlow's finding of persistent deficits in delayed responses of monkeys following frontal lesions, and loss .in discrimination following posterior lesions, illustrates the kind of work that supplies a link between animal and human studies. The question of underlying factors is also involved, since Harlow found that despite clearly localized deficits no task-defined function was completely destroyed by any lesion. "Our theoretical position," Harlow says, "is that, although no specific intellectual function is localized in any single cortical area, the differ ent cortical areas play markedly unequal roles in the mediation of our diverse intellectual processes" (43:252). Harlow is speaking, presumably, of task- defined functions, while Hebb carries the discussion a step further, saying;: 30 The theory that all parts of the cortex contribute equally to problem-solving may mean only that the problem-solving contains so many diverse elements that on the average all parts of the brain con tribute to total performance equally. . . . These unknown components may each be restricted to one part of the brain. (43:256) Teuber, in turn, points out: . . . the problem of localizing psychologic func tions in the cerebrum has at least two aspects: it is necessary (1) to know how to subdivide the neural structure, and (2) to know how to identify the functions to be localized. . . . Every change in procedure of testing is bound to lead to a different classification of cerebral function. (43:259) These three compatible but not identical posi tions may be synthesized as follows. Different arrays of tasks lead to different concepts of the behaviors that may be functionally localized in the cerebrum. As tasks become more unitary and better defined, they approximate the fundamental elements of problem-solving behavior, making localization easier, and yet there is much residual overlapping. Theoretically it should be possible to define the behaviors so as to maximize the possibility of finding a discrete neural base for each one, provided, of course, appropriate and sufficiently exact anatomical divisions can be made. The position of the present study is that factor analysis offers a possibility of adequate definition of unitary behaviors. In what quarter greater possi- 31 bilities lie for adequate anatomical specification may be seen below. Teuber (93) reviews two studies illustrating the combined effects of surgical and chemical variables, and another illustrating the combined effects of sur gical and situational manipulations, upon problem solving behavior. Systematic, multivariant inquiry becomes more feasible as more is known about neuro anatomy, neurophysiology, psychopharmacology, and the structure of behavior. Some outstanding contributions have been made by Pribram and his associates. Commenting on the meaning of psychosurgery, Pribram (74) discounts endur ing physiological alterations and attributes improvements in the patient's adjustment to training during a period of temporary lability following operation. Pribram and Fulton (77) criticize the hasty extrapolation to psychosurgery from conclusions based on animal studies and produce evidence that the conclusions were incorrect in the first place. Pribram and Kruger (78) present a systematic approach to anatomical parcellation, fore telling progress through new surgical techniques that make possible the ablation of formerly inaccessible structures. A striking contrast to these forward strides is seen in a devastating comment by Bailey (43:257-259) on the difficulty of selecting the site of lesion in psychosurgery and even of confirming the accuracy of the removal by microscopic examination of the ablated material. In a study that carries localization to fairly fine distinctions, Pribram et al. (79) demonstrate differential deficit in the delayed-response perform ance of baboons consequent to dorsolateral frontal ablation versus ventromedial frontal ablation. Hypotheses to reconcile old and new data are presented in a major article by Pribram (75). Opportunities for much greater surgical control are revealed by Woolsey (108). A remarkable example of progress in laboratory technique is described by Lilly (60). Arrays of stimulating and pick-up elec trodes implanted on the pial surface of a monkey's cortex activate a grid of lights that are photographed in moving pictures. Short-term patterns of neuro- physiological reaction appear, reflecting such behavioral changes as the passage from waking to sleeping. Lilly says: It is to be expected that as wider areas of the CNS are covered, we shall begin to find closer ties between these figures and behavior even to the point of seeing progressive changes in some of the figures as the monkey progresses in a learning situation. (60:98) 33 Wherry's (103) factor analysis of rat maze learning data comes to mind as a possible prototype of behavioral measures needed to show correspondences with the data foreseen by Lilly. Changes in figures of electronic activity and of their paths of travel over the cortex could reflect changes in the factor composition of tasks in a progressive learning situation. The problem of localization resolves logically into the appropriate specification of loci and behaviors that may be supposed to correspond. A general agreement on brain parcellation lies in the future, though much progress is being made in animal studies. Eventually, new information about psychopharmacological agents (3, 81, 82), as well as other still unknown techniques, may facilitate more powerful investigations of locus in the brains of living, ambulant human beings. The intended contribution of the present study is to the more appropriate specification of behaviors. On either side of the physiologico-behavioral relationship, specification has been too gross or inappropriate to allow any clear pattern of correspondence to emerge. The direction of development on both sides is toward analysis of complexities into unities. CHAPTER III BACKGROUND: FACTOR-ANALYTIC STUDIES A general orientation to factor analysis is given by Wolfle (107) and Thurstone (95). Applications to problems in psychology are contemporaneously con sidered by Fruchter (17) and Guilford (27, 28). Factor analysis as a subject in itself cannot be considered here, but its contribution to this problem and to similar problems can be clarified. The present investigation might be called a secondary application of factor analysis. A primary application would be one having, in Thurstone's words, "the purpose of identifying the principle dimensions or categories of mentality" (95:55). In the present study the purpose is not to discover dimensions, but to apply well-studied instruments to a new population of subjects, to see whether already discovered dimen sions of mentality will reappear. Some of the studies mentioned below are both primary and secondary applications at the same time, trying to show differences between populations and define novel dimensions of mentality all in one set of operations. The boot-strap method in science is not without rewards, but factor structures are hard 34 35 enough to interpret and subtle differences hard enough to demonstrate without confounding the two unneces sarily. It is an advantage to have a ready-made array of indicators, factorially defined in previous studies, to apply to this differential problem. The concept of cerebral localization encountered strong competition with the appearance of Lashley's hypothesis of equipotentiality (56). Progressive decrements in maze learning were said to follow pro gressive removal of cortical tissue, without regard to locus. Excepting sensory and motor areas, any cortical tissue could mediate maze-learning abilities adequately. Quantitative loss of ability resulted from quantitative loss of tissue. Because Lashley's hypothesis became an important issue in psychology, it is all the more interesting to note that twelve years later Lashley altered his view radically. Vector analysis [has revealed] functional vari ables which are not those of classical psychology. . . . They do seem, however, to correspond to functions which may be independently lost as a result of localized brain injury .... The dis covery that the various capacities which inde pendently contribute to intellectual performance do correspond to the spatial distribution of cerebral mechanisms represents a step toward the recognition of similar organization in neurological and mental events. (57:468-469) Amplifying Lashley's observation, it may foe said that limited, localized brain lesions may not 36 relate to changes in abilities that are task-defined and still may be involved in the impairment of abili ties that are factor-defined. A brain-damaged patient may, after a suitable recovery period, be able to perform the same tasks he could perform before injury, and yet, as Harvey suggests (45), the factorial composition of those tasks may have changed for him. Practice in the application of one ability could have compensated for the impairment of another ability. Since most of the tasks that have been used to study intellectual consequences of brain lesions are fac torially complex, this kind of compensation may have obscured the impairment more often than not. The hypothesis of equipotentiality implies that a surviving portion of the brain takes over task- defined ability Z from an injured portion. This raises the question of how an ability migrates from one cerebral locus to another. Harvey's hypothesis, on the other hand, implies that the destroyed tissue has helped to mediate factor-defined ability X, the surviving tissue mediates factor-defined ability Y, and both X and Y are components of task Z. Once the tissue is destroyed, ability X is impaired, and task Z must be accomplished by a greater use of ability Y. Practice should bring about this realignment of 37 abilitieso The period of retraining, during which some decrement might be noted in the performance of task Z, could coincide with the period of recovery from operation or injury. The temporary decrement could be attributed to a temporary physiological dis equilibrium, as it has been in the majority of studies reporting a temporary, but no permanent, loss of per formance. Discussing Lashley's maze-running data and the equipotentiality hypothesis, Harvey says: The task of maze-learning would appear, on inspec tion, to be a task that factorially is quite complex. Thus, perhaps a good deal of visualiza tion is involved, perhaps one or more of the memory factors, etc. Let us assume that there is localization of the regional sort, but within each region there is a punctiform localization corresponding to . . . factors. The extirpation of cortical tissue at a given point may involve the knocking out of a portion or all of the factor which stems from the site. However, it is possible that the factor reduction may only slightly affect the end performance of the multi-factor task . . . . We might extend this to the contention that it is possible to adequately perform on a given task using somewhat different abilities. Thus, in the Lashley study, the reduction of a single factor will perhaps insignificantly affect the total. The reduction of an additional factor, by enlarging the extirpation, might produce a somewhat greater effect upon total performance because we have eliminated the use of another component ability, and so on. Actually, then, as the extirpation becomes larger, the task efficiency is reduced, erroneously presenting a picture of equipotenti ality , while we have, in effect, punctiform localization. (45:42-43) 38 Bringing the issue of localization versus mass action up to date, Orbach (72) cautions that the utility of factor analysis in defining functions that may be cortically localized or focalized has yet to be demonstrated. He emphasizes, however, that the major problem is the definition of unitary functions, and he foretells important progress in neuropsycho logical theory consequent to their definition. Klebanoff et al. (55) mention an early study by Eysenck in which data from senile dements yielded a general factor and three other factors related to speed, memory, and physical strength, as dimensions of deterioration of mental function concomitant with diffuse brain deterioration. A major work, by Halstead, is one of two direct precursors of the present study. Halstead colorfully describes his extensive investigation as "a twelve- year search in a psychological laboratory for 'a man on horseback'" (41:3). The figure is Freud's, the man the ego, the horse the id. Halstead attempts to answer the question: "What is the basic plan or structure of the ego, and whence come its forces?" (41:3). The degree to which the answer succeeds is succinctly evaluated by Thorndike (94); but even though the question is too big, Halstead reports 39 provocative evidence of physiologico-behavioral relations. In the nonfactor-analytic portion of his study, Halstead applies a battery of nine psychological tests, yielding ten scores. Sensory, perceptual, motor, memory, spatial, and abstraction tasks are included. A frontal lobectomy group, a nonfrontal lobectomy group, and a control group are compared. Most of the subjects were not psychiatric patients. A comparison of group mean scores shows that all ten scores differentiate controls from frontal lobectomies; five scores differentiate controls from nonfrontal lobectomies; and seven scores differentiate frontal from nonfrontal lobectomies. All of these differences are statistically significant. In addition to the lobectomized patients, eight patients were tested before and after prefrontal lobotomy (42). The level of performance was com parable to that of the lobectomized patients, but the lobotomy itself seems not to have made much difference. Most of the deficit was attributed to the existing mental illness. Halstead's successes in differentiat ing patient groups appear to rest chiefly on his use of relatively unimpaired patients, rather than on a uniquely appropriate test battery or method of analysis. 40 The factor-analytic portion of Halstead's study (41) is of such a design that relationships between specific cortical lesions and specific impairments of ability are not elucidated. On inspection, however, the results give the impression that the test of memory for shapes may be more sensitive to frontal injury than the other tests in the battery. Despite any shortcomings, Halstead's study marks a turning point in the development of research on the psychological consequences of brain lesions and ablations. In the first Columbia-Greystone study, it is reported that "tests which had shown rather high correlation before operation may decline in their relation after operation .... Tests which show no preoperative correlation may show one after operation" (67:176). A factor analysis is reported, though it is admitted that the number of subjects is too small for such a procedure. We are reminded: Factor analysis is not a method which leads to certain and unique results. The selection of tests and the mathematical processes involved provide opportunities, conscious and unconscious, for finding what the preconceptions of the factor analyst may dictate. Consequently we have used it sparingly in this study. (67:177) One of the uses involves the capricious interpretation of a factor structure derived from a small, incidental 41 battery of complex, factorially undefined instruments; such a use has been held to be a misapplication of factor analysis (22). Another incidental factor analysis is reported by Tow (97), who analyzed difference scores of 36 subjects on twelve indicators taken before and after lobotomy. A general factor of loss of intellectual ability was found. In a well-designed study, Lorr ejt al. explore factors of personality change in lobotomized schizo phrenics. An experimental and a control group, each of 125 subjects, are compared for change on eleven factor scores deriving from a previous study. The comparison is by means of two separate factor analyses of the correlations among changes. Clinical personnel rated all subjects on eighty-one factorially defined rating scales, preoperatively and three months post- operatively, to furnish the data. Four closely similar factors of change were iden tified in the two analyses. The schizophrenic processes tentatively identified were reduced social withdrawal with motor disturbances, reduced schizophrenic excitement, reduced grandiose bel ligerence, and reduced disorganization and dis tortion of thinking. These findings strongly suggest that although lobotomy results in the improvement of many patients with chronic schizo phrenia, the nature of the process of improvement in schizophrenia following lobotomy does not differ greatly from that which may occur without lobotomy. (61:43) 42 In the study most closely related to the present one, Harvey (45) applied factorially defined instruments to groups of normal and defective children and to children with exogenous and endogenous brain damage, attempting to demonstrate group differences in the factorial composition of intellectual tasks independent of differences in mean task scores. Thurstone*s Primary Mental Abilities tests were used with carefully selected subjects. Some suggestive trends developed. No distinction between the endogenous and exoge nous brain-injured could be shown on the basis of test scores alone, and the numbers of subjects were too small to support a comparison of intercorrelations among tests. In the normal-versus-defective comparison, however, correlations were possible, and some signifi cant differences in test intercorrelations appeared between the two groups. The differences could not have been predicted from the mean test scores. The evidence, though only suggestive, tends to support Harvey's contention that intellectual differ ences among such groups are best shown through cor relation techniques and cannot be expected to appear in simple comparisons of means, because of the factorial complexity of tests— even factorially developed tests. 43 Harvey could not get enough subjects to carry out his factor-analytic comparison of three groups of 200 subjects. In addition to a normal control group, there were to have been a group of exogenous brain- injured, selected according to neurological criteria supported by electroencephalographic and pneumoencepha- lographic information, and a group of endogenous brain-injured, restricted to cases of Mongolism and microcephaly. A separate factor analysis was planned for each group. In principle, the present study follows the line laid down by Harvey, comparing patients with damaged to patients with relatively unimpaired brains, applying a battery of factor-defined instruments, and resorting to factor analysis as the mathematical treatment most likely to reveal group differences that might not be demonstrable by a comparison of mean scores. Harvey points out some implications of inter group differences in factor composition of tests, should they be found, for diagnosis and treatment. Diagnostically, a battery could be selected to maximize group differences by means of a regression equation. In treatment, alternate plans of training surviving factor-defined abilities and of attempting to restore impaired factor-defined abilities could be tried out. 44 Harvey also suggests a simultaneous investiga tion of a general intelligence test, the Stanford-Binet, in order to develop a regression equation for the sub tests, maximizing the power of the Stanford-Binet to differentiate the groups under study. In the present investigation, a simultaneous study of the Wechsler- Bellevue Scale is undertaken, though it is not a factor-analytic study and has primarily the purpose of reproducing the kind of data that historically have failed to differentiate lobotomized from unlobotomized schizophrenics, so that the comparison of factor-defined and task-defined tests may be immediate. Before going on to a detailed description of the present research plan, some references must be made to studies of the factorial composition of the Wechsler- Bellevue Scale (99, 100), since it is the contention here that one explanation of that test's failure to differentiate the groups in question is in the uninter preted factorial complexity of its subtest and IQ scores. The best factor-analytic study of the Wechsler- Bellevue Scale to come to the attention of this writer is by Davis (11), who concludes that the subtests in general are factorially complex. Even though reference variables are used, he finds that many of the subtests 45 still have considerable specific variance. At least five of Thurstone's seven primary mental abilities are represented in the Scale. The Vocabulary, Arith metic, and Picture Arrangement subtests are the least complex of the subtests, in Davis's solution. Davis (10) analyzed twenty-six variables, including the Wechsler subtests. The subjects were normal boys and girls, predominantly in the thirteen- to-sixteen-year age range. No correlation was based on fewer than 154 subjects. Three of the factors found by Davis are numerical facility, verbal comprehension, and general reasoning. He found that they were well represented, respectively, by the Digit Symbol, Vocabulary, and Arithmetic subtests of the Wechsler-Bellevue Scale. This finding is the basis for the inclusion of these three subtests in the factor-analytic portion of the present study, along with three factor-analytic marker tests for the same factors. Davis (10) reviews factor-analytic studies of the Wechsler-Bellevue Scale by Balinsky, 1941, Goldfarb, 1941, Simkin, 1951, Cohen, 1951, and Birren, 1951. Only the Goldfarb study included reference variables, but Goldfarb extracted only three factors, did not rotate, and attempted to interpret only two of the factors, as a general and an experience factor. Cohen (8) adds two references, to studies by Hammer, 1950, and Hover, 1950. Guertin et al. (20) refer to studies by Alderdice and Butler, 1952, and Wheeler, 1950. Wechsler (101) adds a study by Gault, 1954, but says that the Davis study (11) is the only one to use reference variables. CHAPTER IV THE PROBLEM The main purpose of this investigation is to demonstrate factor-analytic measurement of behaviors in a physiologico-behavioral problem. The presence or absence of a prefrontal lobotomy constitutes the independent, physiological variable. Factor-defined intellectual abilities are the dependent, behavioral variables. A structure of intellectual factors, derived from a group of lobotomized subjects, is compared with a structure, based on the same tests, but derived from a comparable group of nonlobotomized subjects„ In addition, two kinds of tests are compared: (a) the complex, task-defined intelligence test, in this case the Wechsler-Bellevue Intelligence Scale, Form I, its subtests and IQ scores (99, 100), and (b) a battery of relatively univocal, factor-defined tests, relating to the structure-of-intellect model (28, 38). Group mean score comparisons are made and statistically evaluated, and the comparative ability of the two kinds of tests to differentiate the groups is observed. 47 48 The majority of investigators who have studied intellectual consequences of lobotomy in schizophrenia have reported neither loss of ability nor inferiority of lobotomized to unlobotomized patients (16, 42, 50, 55, 59, 68, 85, 106). Some have depreciated the role of the frontal lobes in intellect (53, 86, 112). Some have suggested that there may be gains after lobotomy or superiority of lobotomized to unlobotomized patients (55, 58, 65, 86)„ In keeping with this back ground, the following hypothesis is tested: Hypothesis: the unlobotomized control group will have mean scores on intellectual indicators that are not different from or are lower than those of the lobotomized experimental group. Three stages are observed in testing the hypoth esis. Though confidence is felt in the comparability of the groups at the first stage, the second and third stages have the effect of increasing that comparability. At the third stage, the comparison between task-defined and factor-defined tests is emphasized by removing intergroup differences on the task-defined variables and observing whether intergroup differences on the factor-defined variables are independent of this manipulation. CHAPTER V PROCEDURES Description of the Tests The thirteen tests of the main battery are listed below. Three of them provide two scores, to make sixteen test variables in all. Unpublished tests bear their code numbers from the Studies of Aptitudes of High-Level Personnel (29, 30, 32-39, 62, 66)o For each test, a brief description of the task and scoring procedure, a sample item (for some tests), the number of parts and items per part, the working time, and the strueture~of-intellect nomenclature of the expected leading factor loading (38), are given. When two working times appear, the one in brackets is the original time limit, before modification in the present study. All tests are pencil-and-paper test s exc ept the last two, in which the subject responds orally. In previous studies, Ship Destination and Social Situations were used with machine-scoring answer sheets. To decrease their difficulty and factorial complexity for abnormal subjects, a different format was adopted for them, and similarly for the Letter 49 ies test. Numerical Operations— Guilford-Zimmerman Aptitude Survey, Part III, Form A (40). Choose one of six numbers that is the correct sum, difference, or product for a given numerical problem. One part, 132 items, 8 minutes working time. Score: number right plus 1/6 number of omitted items. Factor: NSI, numerical facility. Verbal Comprehension~-Guilford-Zimmerman Aptitude Survey, Part I, Form A (40). Select one of five words that is similar in meaning to a given word. One part, 72 items, L25) 12 minutes working time. Score: number right plus 1/5 number of omitted items. Factor: CMU, verbal comprehension. Ship Destination Test (5). Find the distance from ship to port, considering the influence of several variables, such as wind velocity and direction. Circle 1, 2, 3, 4, or 5, as the correct answer, on the answer sheet. One part, 48 items, 15 minutes working time. Score: number right plus 1/5 number of omitted items. Factor: CMS, general reasoning. 51 4-5. Letter Series— an adaptation of Thurstone's PMA Reasoning subtest (96). Circle one of six letters as the next one to follow a given series. Sample item: abababab abcdef Answer: a One part, 30 items, [6.1 12 minutes working time. Score: number right plus 1/6 number of omitted items. Variable 4, odd items; variable 5, even items. Factor: CSS, symbolic patterns (39). 6. Unusual Details— EMS 01B. Write down two incongruities appearing in sketches of common scenes. Sample answers: Lamp cannot give light when cord is not plugged in. Bottom half of clock dial has numbers reversed. Two parts, 16 items per part, L4 J 8 minutes per part. Score: number right. Factor: EMS, experiential evaluation. 7. Social Situations— EP03A0 For a given situation, select the action leading to the most desirable consequences. Circle the right letter: A, B, 52 C, or D. Sample item: You are on a weekend trip with a group of friends. Most of them would prefer spending the day hunt ing, but you would prefer to go fishing. You should: A. Go hunting with them. B. Tell them to go hunting while you go fishing. C. Try to convince them that they will have a better time fishing D. Offer to toss a coin to decide whether the whole group goes hunting or fishing. Answer: A Two parts, 15 items per part, 5 minutes per part Score: number right plus 1/4 number of omitted items. Factor: EMS, experiential evaluation. Seeing Problems— CS06A. Write as many as five problems arising from the presence of a given object. Sample object: Candle. Answers: 1. How to light it. 2. Keeping it from falling over. 53 3. Keeping it from flickering. 40 How long will it burn. 5. What to do with the drippings. Two parts, 6 items per part, [4 3 8 minutes per part. Score: number of appropriate problems listed. Variable 8, Part I. Variable 9, Part II. Factor: EMI, sensitivity to problems. .10-11. Ideational Fluency (6). List objects that belong to specified classes. Sample item: Name FLUIDS that will BURN. Answers: gasoline kerosene hydrogen alcohol Four parts, 1 item per part, 3 minutes per part. Score: number of acceptable responses. Variable 10, Parts I and II. Variable 11, Parts III and IV. Factor: DMU, ideational fluency. 120 Brick Uses (shifts)— CF04A. List different uses for a brick, One part, one item, 10 minutes working time. 54 Score: number of times the subject shifts from one kind of use to another. Factor: DMC, semantic spontaneous flexibility. 13. Alternative Uses— experimental version of Alternate Uses (31). List different unusual uses for common objects. Sample item: Given a newspaper (used for reading), you might think of the following other uses for a newspaper: 1. To start a fire. 2 To wrap garbage in. 3. To swat flies. 4. Stuffing to pack boxes. 5. To line drawers or shelves. 6 » To make up a kidnap note. Two parts, 4 items per part, 6 minutes per part. Score: number of different acceptable uses. Factor: DMC, semantic spontaneous flexibility. 14. Digit Symbol— Wechsler-Bellevue subtest (99, 100). Write symbols corresponding to digits, according to a key. One part, 67 items, 1 1/2 minutes working time. Score: number right plus 1/2 number reversed. Factor: NSI, numerical facility. 55 15. Vocabulary— Wechsler-Bellevue subtest (99, 100). Orally define a list of words orally presented. One part, 42 items. Score: sum of unit and half credits, according to standards in the manual (99). Factor: CMU, verbal comprehension. 16. Arithmetic— Wechsler-Bellevue subtest (99, 100). Without a pencil, solve orally- or visually- presented problems of the type: How many hours will it take a man to walk 32 miles at the rate of four miles an hour? One part, 10 items, 15 to 120 seconds per item. Score: number right plus time credits on items 9 and 10, according to the manual (99). Factor: CMS, general reasoning. Selection of the above tests was made for a number of reasons. It was thought desirable to include some much-studied factors, such as numerical facility, verbal comprehension, and general reasoning. A second desideratum was to have at least one factor from each of the four thinking divisions of the structure-of- intellect model: cognition (CMU, CMS, CSS), divergent production (DMU, DMC), convergent production (NSI), and evaluation (EMI, EMS)„ There was particular interest in productivity, since it has been suggested 56 that productiveness (16) and flexibility (9) may be dependent upon the frontal lobes. A third requirement was to favor the more univocal tests, so that the interpretation of the resulting factors might be less ambiguous. A fourth requirement was to have tests of suitable length and difficulty for research with schizophrenic subjects. Economy of testing time was achieved by making three Wechsler-Bellevue subtests do double duty, serving both in the factorial and Wechsler-Bellevue portions of the study. The Wechsler-Bellevue was preferred to the newer Wechsler Adult Intelligence Scale (101), because it had been more studied, had been adequately factor-analyzed, was a source of early scores on the subjects for possible supplementary study, and was not so likely to have been recently administered to the subjects. The scoring of tests in separate parts and the intercorrelation of those parts results either in the appearance of doublet factors or in the inflation of factor loadings by the confounding of common and specific factor variance (66). Some risk of restric tion of interpretability of the factors was thought to be justified by the larger number of factors thus introduced. The same argument applies to the limitation of marker variables to two per factor, sacrificing the overdetermination of factors that is desirable in primary applications of factor analysis. The Sub.jects One hundred fifty lobotomized and 150 unlobot- omized schizophrenic patients were selected from the rolls of eleven hospitals in California. More than half were veterans, from Brentwood, Palo Alto, and Sepulveda Veterans Administration Hospitals. The rest were from Agnews, Atascadero, Camarillo, Mendocino, Metropolitan, Napa, Patton, and Stockton State Hos pitals . Experimental group. The 150 lobotomized sub jects were chosen from a pool of over 800 available patients, to satisfy seven criteria. Ail had been given a bilateral prefrontal lobotomy, and were not known to have suffered any other brain damage. All had a diagnosis, before lobotomy, of paranoid, hebe phrenic , catatonic, simple, undifferentiated, or mixed schizophrenia, with no other diagnosis to complicate the picture either before or after operation, except for a diagnosis, in 33 cases, of epilepsy following lobotomy. No subject was over sixty years of age. 58 All spoke English as their most used language and had finished the eighth grade in school. The postoperative interval, or remoteness of operation, varied from five to fifteen years, with a mean of 8.7 years and a standard deviation of 2.2 years. Control group. The 150 unlobotomized subjects were chosen to match the experimental group in age, highest school grade completed, duration of illness, sex, race, diagnostic subgroup, veteran status, whether or not on a tranquilizer at the time of testing, and hospital source. Comparability of the groups and subgroups. Group matching on the continuous variables, age, schooling, and duration of illness, is shown in Table 1. The differences in group means on all three variables are statistically nonsignificant and are very small. The fact that the control subjects are more variable in duration of illness than the lobot omized subjects is attributable to the fact that all lobotomies were performed within an eleven-year period, tending to constrict the variability of the lobotomized subjects on chronological variables. This effect is increased in the veteran sample, shown in Table 2. The interpretation of Tables 1 and 2 is TABLE 1 GROUP MATCHING ON CONTINUOUS VARIABLES 150 Experimentals 150 Controls Ratiosa Variable Mean SD Mean SD F t a. Age 41.86 6.48 42.09 6.75 1o 09 0.30 b, Schooling'3 11.82 i. 70 11.71 1 o 93 1.29 -0.52 c. Illness0 16.59 1.23 16.63 5.40 1.62** 0.07 F-ratio has two-tailed test of significance; t-ratio has one-tailed test. Two stars indicate .01 level of significance or better. 1 ^ Highest grade completed. Q Duration of illness. TABLE 2 SUBGROUP MATCHING ON CONTINUOUS VARIABLES 69 d Experimentals 72 Controls^ Ratiosa Variable Mean SD Mean SD F t Ci 9 Age 40.03 5.35 40.19 6.53 1.49* 0.16 b . Schooling*3 11.65 1.73 11.51 1.93 1.25 -0.05 c. Illness0 16.17 3 0 62 15.97 5.06 1.95** -0.27 50 ^ Experimentals 50 Controls^ C 2 . o Age 39.90 4. 90 39.88 6.53 1.78* b. b Schooling 11.68 1.78 11.50 2.08 1.36 -0.46 c. Illness0 16.08 3.62 15.82 0 • 07 1.96* -0.29 z . WB FS IQe 97.98 12.70 98.28 13. 53 1.13 ... aF-ratio has two'-tailed test of significance; t-ratio has one-tailed test. One star indicates .05 level; two stars indicate .01 level of significance or better. I d Highest grade completed. c Duration of illness. ^White veterans. eVV'echsler-Bellevue Scale, Form I, Full Scale IQ. 61 that there are no meaningful d fferences in age, schooling, or duration of illness, either between the main groups, or between subgroups. Group matching on category variables is shown in Table 3. The first two columns compare the main groups; the last four columns compare subgroups. Matching in all categories is very close, with two minor exceptions. More lobotomized subjects are diagnosed as catatonic; more controls are diagnosed as undifferentiated. If any bias is introduced by this discrepancy, it probably tends to reduce test score differences between lobotomized and control subjects; for in the last two columns, where group mean Wechsler-Bellevue IQ's have been equated, the discrepancy is slightly increased. Matching for race is not shown in the tables. Included in each group of 150 subjects were two Negroes and three Mexicans. All others were Caucasians, except for one American Indian in the experimental group. All subjects in the subgroups were Caucasians. All feasible steps were taken to increase the comparability of the lobotomized and control groups. Ideally, the controls would have been schizophrenic patients who had been chosen for TABLE 3 CATEGORICAL GROUP AND SUBGROUP MATCHING a v - ' S . L -egory 150 E° 150 C 69 E 72 C 50 E 50 C d. Sex— Male .53 .53 .87 .88 .82 .82 e. Paranoid .47 .47 .52 .53 . 54 .54 Hebephrenic .17 . 15 .12 .11 .12 .12 u. Catatonic .22 .13 .19 .13 .26 .14 Vo Simple .01 . 03 .00 .03 .00 .00 Wo Undifferentiated .07 .16 .13 .18 . 06 .16 "V * J ' t G Mixed . 05 .05 .04 .03 o 02 .04 g « Veteran . 54 e 54 1.00 1.00 1.00 1.00 h9 Tranquilizer .74 .77 .88 .90 . 86 .86 y* Hospital0 .94 « * • .97 • • • .84 • • • cat q Proportions egories. of groups falling into criterion or descriptive b E, Experimentals; C, Controls. Groups 69E, 72C, 50E, and 50C are white veterans. cProportion of experimental group that is matched by having controls from the same hospital. cn ro 63 lobotomy and then had been selected by the investigator to act as controls instead of having the operation. Such a design, of course, could never be carried out, due to extra-experimental considerations. An approxi mation to this ideal was to find subjects who had been recommended for a lobotomy at a medical staff confer ence and then had not received the operation because of some event extrinsic to the individual diagnosis and plan of treatment. Such events were the refusal of relatives to grant permission for the operation, transfer of the patient, or a change of policy with respect to lobotomy. At every hospital where controls were obtained, a search was made for such subjects; fourteen were found. This control-group category, Lobotomy Recom mended , was correlated with each of sixteen test variables, to see whether there were a tendency for the recommended subjects to do less well than the other controls. The correlation coefficients appear in Row k of Table 7. None is large enough to be statistically significant; eleven are negative and five are positive, while the chance expectation is eight and eight. The interpretation of these results is that the test performance of the control group would not have been of lower quality if all controls 64 had been previously recommended for lobotomy. Support ing this interpretation is the fact that in the veteran sample, where six out of 72 subjects were recommended for lobotomy, the mean Wechsler-Bellevue Full Scale IQ of the six was 3.5 IQ points higher than that of the other 66 subjects, a nonsignificant difference. Not shown in any table is the fact that an effort was made to obtain control subjects from the same wards as lobotomized subjects, and in the same proportions. To a large extent, this effort was successful. Bias in terms of inter-ward variables was thus minimized, and it is felt that no advantage was given to the control group as a result of inter ward variables. Another aspect of the attempt to avoid a bias in favor of the control group was the inclusion of nine lobotomized veterans who were not, like all controls, residents in a hospital at the time of testing. Two were residents in the Domiciliary, five were in Family Care, one was on Trial Visit, and one was discharged. The discharged patient was still mildly delusional and hallucinated. Two of the subjects in Family Care are eliminated from the sub group of 50 lobotomized subjects. 65 The comparability of experimentals and controls is increased in the intermediate subsample, of 69 experimentals and 72 controls, by restricting the sub sample to white veterans. The elimination of non whites had the actual effect of reducing differences in mean test scores between experimentals and controls. Coincidentally, a small number of subjects were lost because of incomplete Wechsler-Bellevue data. No systematic trends in the reasons for the loss appeared; and this restriction likewise had the actual effect of reducing mean score differences between groups. In the small subsample, intergroup comparability is increased further, by equating subgroup means and variabilities on the Wechsler-Bellevue, Form I, Full Scale IQ for 50 experimentals and 50 controls, drawn from the intermediate subsample. As can be seen in Tables 1, 2, and 3 (pages 59, 60, and 62), each step of this progressive matching procedure was achieved without sacrificing any of the previous matching. No effort was made to measure motivation in the two groups apart from their test performances, where it is inseparable from ability. The investi gator received no reports of differential motivation from the examiners. The only evidence bearing on this point is the fact that, on some tests, mean score 66 differences were extremely small, which argues against a general difference in motivation. No attempt was made to match the groups in terms of their histories of shock therapies. Notation in clinical records was not consistent enough or complete enough to justify assigning a numerical value for each subject. It was assumed that two groups of patients of such comparability and chronicity would have had the benefit of the various forms of shock therapy in about equal measure„ A possible difference of policy on shock therapy from hospital to hospital was con trolled by a high degree of hospital matching (see Table 3, page 62). No subject had had shock treatment for six months preceding testing. In a large majority of cases, the subjects had not had shock treatment for several years. The Pilot Project Before undertaking to test 300 subjects, it was thought advisable to find out just how difficult the tests were for chronic schizophrenic patients and whether any alterations in the time limits would be desirable. Accordingly, the main battery, plus the Logical Reasoning test (48) was administered to 27 patients. Two were unlobotomized; nine had 67 prefrontal, and 16 had transorbital, lobotoinies. The 11 patients in the first two classes are included in the main study. Working times of all tests, except the three Wechsler-Bellevue subtests, were modified. Scores were recorded for the original time limits and the modified time limits. A frequency distribution was plotted for each test variable at each time limit. It was found that extending the working time made very little difference in the shape of the dis tribution or in the number of zero scores for Numerical Operations, Ship Destination, Social Situations, Ideational Fluency, Brick Uses (shifts), or Alternative Uses. Reducing the working time for Verbal Compre hension made very little difference, and so the shorter time was adopted. On the other hand, doubling the time limits on Unusual Details, Seeing Problems, and Letter Series improved the distributions and reduced the number of zero scores, and so the longer times were adopted. Logical Reasoning was too difficult even for experi mental use, and was dropped. 68 Administration and Scoring; of Tests Nonveterans were tested between November, 1959, and October, 1960; veterans, between March, 1960, and January, 1961. All subjects were tested individually, in two or more sessions. Each subject had about three hours of testing, or four hours if the entire Wechsler- Bellevue were given. Twenty-one examiners, including the investigator, administered the tests. Eight were clinical psychology trainees; five were hospital staff psychologists; and three were school psychologists. Of the remaining five, two had had graduate training in testing; two were regularly engaged in administering tests to hos pitalized patients; and the fifth had had both graduate training and hospital testing experience. The geographical scatter of the subjects and the extended testing schedules made it difficult to match the groups on examiners. However, 0.58 of the experi mental subjects have controls who were tested by the same examiner. The experience of the examiners with the battery was quite comparable for the two groups. Serial administration effects were controlled by the application of six counterbalanced test orders, dis tributed about equally among the subjects. Though 69 there was a tendency for the iobotomized subjects to be tested earlier than the controls, there was a great deal of overlap. The Wechsler-Bellevue subtests were administered according to the manual (99). When the balance of the Wechsler was given, it followed the main battery. The factor-analytic tests were administered with the help of a special manual prepared for the project, providing a procedure for adapting the standard group instruc tions to individual administration. The key passage was as follows: In administering these tests individually, the examiner should use the instructions judiciously, as a guide for himself and as information for the subject. Help the subject to understand the task as fully as possible. Do not read the instruc tions verbatim if the subject has any difficulty following them, but explain the tasks making use of the pertinent instructional information. It is not necessary to tell the subject how long the test is or how many parts it has. It is necessary to observe him closely to make sure that he is complying with the instructions. Give help, if needed, during the instruction period. During the test proper, do not give help, but encourage the subject to keep going if his interest flags. Very few anomalies in test administration came to the attention of the investigator, who scored all of the tests. The test batteries were scored as received, which led to a great deal of overlap between the 70 experimental and control groups. No blind scoring procedure was used, since most of the scoring was completely objective. Four tests, Seeing Problems, Ideational Fluency, Brick Uses (shifts), and Alter native Uses, required the scorer to evaluate the responses to a degree that might nave permitted a scoring bias to affect the results. Two of these tests differentiated the groups at a high level of significance; the other two did not differentiate them,, Any scoring bias that may have existed appears not to have favored either group systematically. Treatment of Data At every stage of the treatment of data, and of the factor analyses, identical or logically equivalent mathematical procedures were applied in the two groups, in order to minimize spurious inter group differences. Continuous variable data. Distribution of the sixteen test scores and of age, schooling, duration of illness, and remoteness of operation were plotted graphically and inspected for unimodality, symmetry, and amount of kurtosis. Arithmetic, Digit Symbol, and Ideational Fluency were found to be sufficiently 71 isokurtic to justify the use of raw scores in the computation of Pearson product-moment correlations. Evaluations of the other variables were as follows. Bimodal: remoteness of operation. Leptokurtic: schooling; the mode was twelve years. Mild positive skew: age, duration of illness, Unusual Details, and Social Situations. Moderate positive skew: Numerical Operations, Verbal Comprehension, Seeing Problems I. Severe positive skew: Ship Destination, Letter Series, Seeing Problems II, Brick Uses (shifts), and Alterna tive Uses. Mild negative skew: Vocabulary. Variables that were bimodal, leptokurtic, or mildly or moderately skewed were normalized graphically (23). Severely skewed variables were dichotomized as near as possible to the median (23), subject to the provision that a zero score would not be included in the upper category. This provision resulted in some nonmedian splits, the most extreme of which was a 0.61-0.39 split. Means, standard deviations, and reliability estimates of the test variables are shown in Tables 4 and 5. Category data. Sex, veteran status, paranoid, hebephrenic, tranquilizer status, history of seizures, TABLE 4 MEANS, STANDARD DEVIATIONS, AND RELIABILITIES OF SCORES EXPERIMENTAL GROUP Variable M SD ► 1 r t - i. Numerical Operations 41.89 12.51 .53 2. Verbal Comprehension 29.47 11.57 .79 •3. Ship Destination 14.28 5. 53 .85. 4. Letter Series— odd 4.03 3.11 .96 5. Letter Series— even 4.31 3.23 .88 6. Unusual Details 11.02 6.29 .65 7. Social Situations 11.15 3.22 .65 8. Seeing Problems— I 4.75 4.22 .89 9. Seeing Problems— II 3.25 3.94 .82 10. Ideational Fluency— X-II 13.38 6.36 .72 11. Ideational Fluency— III-IV 17.05 8.37 .74 12. Brick Uses— shifts 2.63 3.09 .74 13. Alternative Uses 4.34 4.20 .69 14. Digit Symbol 29.09 9.45 .60 15. Vocabulary 20.57 6.88 .85 16. Arithmetic 5.48 2.73 .64 aCommunalities as lower-bound estimates. 1) From Control Group. Obtained h appears an overestimate due to confounding of common and specific factor variances. TABLE 5 MEANS, STANDARD DEVIATIONS, AND RELIABILITIES OF SCORES CONTROL GROUP Variable M SD rttS 1. Numerical Operations 47.49 15.57 .82 2. Verbal Comprehension 32.91 12.27 .80 3. Ship Destination 16.85 7 o 3 o .79 4. Letter Series— odd 5.42 3.48 .96 5. Letter Series— even 5.74 3.16 .98 6 o Unusual Details 12.20 7.32 .79 7. Social Situations 12.27 3.85 .47 8. Seeing Problems— I 6.80 5.19 .86 9. Seeing Problems— II 5.46 4.84 .98 10. Ideational Fluency— I-II 13.87 6.42 .70 11. Ideational Fluency— III-IV 18.67 8.67 .77 12. Brick Uses— shifts 2.99 3.80 .75 13. Alternative Uses 6.15 o*45 .66 14. Digit Symbol 29.15 10.52 .60 15. Vocabulary 21 o 62 6.47 .80 16. Arithmetic 6.09 2.79 .69 aCommunalities as lower-bound estimates. 74 and recommendation for lobotomy were category variables for correlation. The other four diagnostic subcate gories were not correlated, because of insufficient numbers or unequal proportions in the two groups (cf. Table 3, page 62). Intercorrelations. Pearson product-moment correlations were computed for all pairs of variables on an International Business Machines 709 Computer. Correlations involving dichotomized continuous vari ables were corrected for coarse grouping (23). In three cases, the limitations for the application of the correction were exceeded, and tetrachoric cor relations were computed for both groups (23) and corrected for nonmedian dichotomization (73). In the original correlation matrices, an addi tional variable was included, Brick Uses (fluency) (66), which is not independent of Brick Uses (shifts), and which was dropped because it seemed unlikely to contribute to the clarity of the eventual solution. The final matrices of intercorrelations are presented in Tables 6 and 7. 75 TABLE 6 CORRELATION MATRIX,3 EXPERIMENTAL GROUP Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 a. Age -11 15 -04 -19 -17 -16 05 00 -09 -15 -08 -09 02 -15 18 07 b. Schooling (highest grade) 16 29 17 34 32 17 21 08 -01 14 12 01 04 16 27 20 c. Illness (duration) -01 -02 -17 -15 -08 -13 -19 -02 -15 -13 00 04 01 -23 -02 -06 d. Sex (male) -01 -07 24 09 04 05 -14 -03 01 -08 -16 -08 04 -12 05 24 e. Paranoid -04 09 01 00 01 00 13 -09 -09 -03 00 06 11 -11 21 20 f. Hebephrenic -02 -05 -16 01 05 -02 -13 04 04 -05 -02 -05 -23 -06 -14 -16 g° Veteran -09 -11 13 00 -04 13 -15 -03 02 -04 -11 -03 03 -10 06 18 h. Tranquilizer -16 -17 -19 -18 -21 02 -19 -08 03 -06 -14 -15 -06 -04 -08 -01 i. Seizures -08 -06 -13 -04 10 -06 -05 01 -06 -05 -13 -03 -07 -03 -09 02 h Operation (Remoteness) 00 -08 -21 -09 -12 -03 -08 -07 -05 -04 - 0 . 1 01 -07 12 -11 -26 1. Numerical Operations 31 49 45 36 43 39 22 18 33 44 16 ■ ) ' ? u t j 47 30 40 2. Verbal Comprehension 31 45 49 42 54 58 39 28 36 43 32 40 23 78 35 3. Ship Destination 49 45 73 64 59 61 40 23 47 45 25 43 35 54 66 4. Letter Series—odd 45 49 73 93 60 57 46 35 45 41 33 47 35 39 55 5. Letter Series—even 36 42 64 93 52 54 29 18 35 26 20 31 38 33 39 6. Unusual Details 43 54 59 60 52 54 48 33 47 48 40 48 41 58 30 7. Social Situations 39 58 61 57 54 54 40 33 42 42 17 34 42 56 32 8, Seeing Problems—I 22 39 40 46 29 48 10 84 38 41 44 ■ 1 8 28 35 24 9. Seeing Problems—II 18 28 23 35 18 33 33 84 20 26 40 36 29 19 14 10. Ideational Fluency—I-II 33 36 47 45 35 47 42 38 20 69 35 33 48 39 23 11. Ideational Fluency—III-IV 44 43 45 41 26 48 42 41 26 69 47 49 39 43 28 12. Brick Uses—shifts 16 32 25 33 20 40 17 44 40 35 47 64 26 36 20 13. Alternative Uses 23 40 43 47 31 48 34 48 36 33 49 64 14 45 31 14. Digit Symbol 47 23 35 35 38 41 42 28 29 48 39 26 14 20 18 15, Vocabulary 30 78 54 39 33 58 56 35 19 39 43 36 45 20 47 16. Arithmetic 40 35 66 55 39 30 (Continued) 32 24 14 23 28 20 31 18 47 dl)eciial points have been omitted. TABLE 6 (Continued) 76 Variable a b c d e f E h i a. Age 10 38 -18 14 07 -26 -15 -07 06 bo Schooling (highest grade) 10 -02 -13 -02 00 -09 -08 -06 -02 c. Illness (duration) 38 -02 -08 -19 26 -13 -09 02 25 d. Sex (male) -18 -13 -08 03 -05 74 24 04 -35 e, Paranoid 14 -02 -19 03 -42 10 04 -15 -17 f. Hebephrenic 07 00 26 -05 -42 -09 06 -02 09 g. Veteran -26 -09 -13 74 10 -09 37 01 -35 ho Tranquilizer -15 -08 -09 24 04 06 37 -05 -12 i. Seizures -07 -06 02 04 -15 -02 01 -05 -10 j. Operation (Remoteness) 06 -02 25 -35 -17 09 -35 -12 -10 1, Numerical Operations - 1 1 1 6 -01 - 0 1 -04 -02 -09 -16 -08 00 2, Verbal Comprehension 15 29 -02 -07 09 -05 -11 -17 -06 -08 3. Ship Destination -04 17 -17 24 01 -16 13 -19 -13 -21 4, Letter Series—odd -19 34 -15 09 00 01 00 -18 -04 -09 5. Letter Series—even -17 32 -08 04 01 05 -04 -21 10 -12 6. Unusual Details -16 17 -13 05 00 -02 13 02 -06 -03 7. Social Situations 05 21 -19 -14 13 -13 -15 -19 -05 -08 8. Seeing Problems—I 00 08 -02 -03 -09 04 -03 -08 01 -07 9, Seeing Problems—II -09 -01 -15 0 1 -09 04 0 2 03 -06 -05 10. Ideational Fluency—I~II -15 14 -13 -08 -03 -05 -04 -06 -05 -04 11, Ideational Fluency—III-IV -08 1 2 , 00 -16 00 -02 -11 -14 -13 -01 12. Brick Uses—shifts -09 0 1 04 -08 06 -05 -03 -15 -03 01 13. Alternative Uses 02 04 01 04 11 -23 03 -06 -07 -07 14, Digit Symbol -15 16 -23 -12 -11 -06 -10 -04 -03 12 15, Vocabulary 18 27 -02 05 21 -14 06 -08 -09 -11 16. Arithmetic 07 20 -06 24 20 -16 18 -01 02 -26 77 TABLE 7 CORRELATION MATRIX,® CONTROL GROUP Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 a. Age -18 10 -10 -14 -24 -30 -08 -15 02 -03 -03 -08 02 -22 -01 -05 b. Schooling (highest grade) 20 39 42 23 29 23 19 30 27 24 28 30 27 23 41 31 c. Illness (duration) -20 02 -09 -18 -20 -24 -08 -10 00 -03 05 04 -07 -18 -06 -08 d. Sex (male) 16 -09 21 -09 06 10 03 11 -10 00 -01 -03 09 -09 08 37 e. Paranoid 07 10 20 34 26 12 19 05 -08 -03 09 04 09 08 21 14 f. Hebephrenic -03 -13 -18 -14 -13 -15 -14 -15 01 -10 -18 -20 03 01 -21 -05 g. Veteran 24 -03 16 00 17 18 11 11 -02 11 03 -02 08 05 08 24 h. Tranquilizer 07 -09 01 01 09 10 03 01 -12 10 09 -08 11 10 00 -06 k. Lobotomy Recommendation -11 04 04 -02 -03 -15 -06 -04 -01 -06 -08 -12 -14 00 03 01 1. Numerical Operations 43 4 7 57 59 » ' c i Ow 45 13 35 49 38 32 40 64 F j 01 60 2. Verbal Comprehension 43 43 64 48 53 46 40 40 39 43 33 41 39 78 44 3. Ship Destination 47 43 57 68 50 32 42 54 34 37 52 62 26 51 51 4. Letter Series—odd 57 64 57 91 44 36 35 46 27 34 22 32 32 51 44 5. Letter Series—even 59 48 68 91 51 37 41 46 34 45 25 52 41 44 41 6. Unusual Details 62 53 50 44 51 57 55 57 42 46 43 63 55 62 46 7. Social Situations 45 46 32 36 37 57 44 40 38 44 32 40 38 44 31 8, Seeing Problems—I 43 40 42 35 41 55 44 80 38 50 46 57 40 44 50 9. Seeing Problems—II 35 40 54 46 46 57 40 80 2 0 55 60 27 37 32 10. Ideational Fluency—I-II 49 39 34 27 34 42 38 38 2 0 6 8 52 43 42 36 38 11. Ideational Fluency—III-IV 38 43 37 34 45 46 44 50 25 68 53 36 39 41 40 12, Brick Uses—shifts 32 33 52 2 2 25 43 32 46 55 52 53 47 27 34 20 13. Alternative Uses 40 41 62 32 52 63 40 57 60 43 36 47 32 38 28 14. Digit Symbol 64 39 26 32 41 55 38 40 27 42 39 27 32 30 30 15. Vocabulary 37 78 51 51 44 62 44 44 37 36 41 34 38 30 50 16. Arithmetic 60 44 51 44 41 46 (Continued) 31 50 32 38 40 20 28 30 50 aDecimal points hays been omitted. TABLE 7 (Continued) 78 Variable a b c d e f g h k a, Age 06 49 -28 01 14 -32 -24 06 b. Schooling (highest grade) 06 -03 -01 02 -05 -08 -16 01 c. Illness (duration) 49 -03 -06 -14 14 -13 -15 -02 d„ Sex (male) -28 -01 -06 11 -05 74 15 -07 e, Paranoid 01 02 -14 11 -40 10 02 16 f. Hebephrenic 14 -03 14 -05 -40 -05 -07 -07 g„ Veteran -32 -08 -13 74 10 -05 34 -07 h, Tranquilizer -24 -16 -15 15 02 -07 34 01 k. Lobotomy Recommendation 06 01 -02 -07 16 -07 -07 01 1. Numerical Operations -18 2 0 -20 16 07 -03 24 07 -11 2. Verbal Comprehension 10 39 0 2 -09 1 0 -13 -03 -09 04 3. Ship Destination -10 42 -09 21 20 -18 16 01 04 4. Letter Series—odd -14 23 -18 -09 34 -14 00 01 -02 5. Letter Series-even -24 29 -20 06 26 -13 17 09 -03 6. Unusual Details -30 23 -24 10 12 -15 18 10 -15 7. Social Situations -08 19 -08 03 19 -14 11 03 -06 8. Seeing Problems—I -15 30 -10 11 05 -15 11 01 -04 9, Seeing Problems—II 0 2 27 0 0 -10 -08 Oi r i o - 1 2 -01 10. Ideational Fluency—I-II -03 24 -03 0 0 - • 0 3 -10 ii 10 -06 11. Ideational Fluency—III-IV -03 28 05 -01 09 -18 03 09 -08 12. Brick Uses—shifts -08 30 04 -03 04 -20 -02 -08 -12 13. Alternative Uses 0 2 27 -07 09 09 03 08 1 1 -14 14. Digit Symbol -22 23 -18 -09 08 01 05 10 00 15. Vocabulary -01 41 -06 08 21 -21 08 00 03 16. Arithmetic -05 31 -08 37 14 -05 24 -06 01 Statistics 79 At the beginning of this study, it was hoped that a suitable test for evaluating the statistical significance of differences between two factor struc tures would be found. Recent publications (4, 28, 44, 52, 64) suggest that, though progress is being made, no test fulfilling the needs of the present study is yet available. When such a test arrives, it can be applied to the present factor solutions. in the nonfactor-unalytic portion of the study, it was necessary to evaluate intergroup inean-score differences, on tests and continuous matching vari ables. The variance ratio is evaluated in every case with a two-tailed F-test (14), since no prediction was made that one group would be more variable than the other. Group mean scores on continuous matching variables are so close that the t-ratios (23) are obviously nonsignificant. On the test variables, a one-tailed t-test (1, 15) is used, since the direction of the difference was specified in advance. In instances where the F-ratio is significant, the t-ratio has been computed with an estimate of popula tion variance made from both groups (13). It has been pointed out (23) that when popula tion score distributions are suspected of being 80 seriously skewed, the t-tests may not apply. Some of the tests in the battery have skewed score distributions with this sample, and there is reason to suppose that any similar sample would produce similarly skewed dis tributions. The central limit theorem (49), however, supports the use of t when samples are as larg;e as in the present study. CHAPTER VI THE FACTOR ANALYSES Extraction of Factors Using Thurstone's complete centroid method (95), nine factors were extracted from the two correlation matrices of sixteen test variables, by an International Business Machines 650 Computer. The distributions of residual coefficients after the extraction of the ninth factor were leptokurtic about zero and contained no coefficient as great as 0.06 in absolute value. The centroid matrices are presented in Tables 8 and 9. Rotation of Axes Graphic orthogonal rotations were used (109). Primary emphasis was laid on the criteria of simple stract ure and posit i ve mani f o1d (95). In the fina1 rotations, the criterion of psychological meaning was imposed to effect minor improvements in the structures. The experimental centroid matrix was rotated first, and an attempt was made to make the control solution conform to the experimental solution. This attempt was only partially successful; substantial unpredicted factor loadings appeared in the control solution, and attempts to rotate them out of existence grossly 81 TABLE 3 CEiNTRUlD MATRIX , EXPERIMENTAL GROUP V a r i a b l e I IX Ill IV V VI V I I VIII IX ...... h * 1. 55 11 10 31 0 9 -16 - 2 2 05 -16 53 2. 69 22 - 2 0 -14 - 3 7 2 0 - 1 2 - 1 4 -07 79 O 0 77 33 -05 11 14 - 2 5 18 10 04 85 4 A S 82 O X 25 -17 36 11 16 - 1 0 - 0 5 103 5 . 68 o 3 32 - 0 7 30 24 12 - 0 5 07 88 6 » 75 05 05 0 / -11 15 04 16 -15 65 ( o 70 21 11 12 - 24 0 8 07 07 13 65 8 • 67 _4.il 2 G - 2 8 -21 - 2 0 18 - 0 4 - 0 2 89 a 52 —i 5 37 - 3 2 - 2 2 - 2 2 , \ sr — KJO 03 04 82 10. 64 -21 - 1 2 38 05 16 14 - 2 2 13 72 11. 67 -27 — 26 33 06 05 07 - 0 9 -11 74 12 o 54 -43 — 25 - 2 2 19 19 - 1 9 21 - 0 4 74 13. 63 - 23 - 2 5 - 2 4 11 06 09 25 -18 69 14. 51 -03 25 39 04 05 -26 08 18 60 15. 69 2 G - 4 0 -11 -37 07 - 1 0 -04 -06 85 16. 55 — i G-i. -17 - 0 7 19 -39 _li -11 -05 64 Note: decimal points ha\>-e been omitted. TABLE 9 CENTROID MATRIX, CONTROL GROUP variable i II III IV V VI VII VIII IX h2 1. 72 21 10 20 -37 0 9 22 11 04 81 2. 71 20 24 - 2 5 2 0 -18 - 1 8 10 - 1 2 80 3. 72 -31 19 06 2 0 11 26 06 15 79 A "X © 70 -17 55 13 -14 -13 - 26 08 07 96 5c 75 - 2 9 39 28 - 1 5 -20 - 0 5 -15 0 9 98 6 c 79 12 - 1 2 -18 - 1 2 -17 20 07 14 79 7 . 61 17 — 13 -11 -08 -14 - 0 5 -03 11 47 8. 73 -1 7 -31 - 2 5 -18 25 - 1 4 -13 - 11 86 9 e 68 — 51 -26 - 2 9 -17 09 - 1 4 23 -0 2 98 10. 62 28 -24 o 2 23 11 - 0 3 -11 ! O cn 70 i l « 65 21 - 2 2 26 20 07 - 2 4 -23 17 77 12. 58 -11 - 3 5 16 31 17 -06 34 09 75 to 68 - 2 4 — 2o -06 10 -13 23 -05 -06 66 14, 58 28 - 0 9 17 -29 - 1 0 11 13 - 1 5 60 JL O s 70 2 0 23 - 3 5 28 -07 - 0 6 09 07 80 16o 61 15 23 - 0 8 -06 43 1 £ 7 „ L <J -13 08 69 Note: decimal points have been omitted. 84 violated the criterion of simple structure. Confirma tion was sought by applying another rotation method. Analytical rotations, to the varimax criterion (51), were computed on the IBM 709 Computer. The analytical and graphic solutions of the experimental rotation problem were very similar, calling for the same interpretations, though the graphic solution more nearly approached simple structure. The control prob lem, however, had markedly different solutions by the two methods of rotation. The analytical solution was further from simple structure, and it exaggerated and emphasized the unpredicted factor loadings found by the graphic method, to the point of producing some quite different factors. It was concluded that the analytical solutions were confirmatory of, though less easily interpreted than, the graphic solutions. Par ticularly, the suggestion that the two factor structure were not essentially identical seemed confirmed by the analytical solutions. The possibility of computing oblique analytical rotations was considered and rejected. It was felt that oblique rotations would tend to maximize the dif ferences between the two groups. In the absence of statistical evaluation, it was thought that the more conservative orthogonal treatment would support 85 inferences about differences in factor structure better. In the experimental solution, sixty-eight graphic rotations were made; in the control solution, ninety-three. The rotated factor matrices are pre sented in Tables 10 and 11. Interpretation of the Factors For each factor, the test variables having significant loadings, of 0.30 or more (66), are listed in order of magnitude. The notations follow ing the factor loading indicate loadings of 0.30 or more on other factors, when there are any. The experimental group is listed first for each factor, as is indicated vertically at the 1 eft margin. Factor A— NSI— - numerical faci1ity K 14. Digit Symbol .54 CSS .36, DMU .34 x I. Numerical Operations .45 CMS .35 C 14. Digit Symbol .63 o 1. Numerical Operations .61 CMS .41, CSS .33 n 6. Unusual Details .38 EMS .47, DMC .42 It has been suggested (38) that numerical facility may be essentially a memory ability, involv ing long-term memory for symbolic implications, rather 86 TABLE 10 ROTATED FACTOR MATRIX, EXPERIMENTAL GROUP Variable A B C 1 ) E F G 1 1 R h2 1, Numerical Operations 45 04 35 27 00 18 16 10 23 52 2. Verbal Comprehension 11 74 09 24 15 06 16 09 27 76 3, Ship Destination 08 19 56 41 08 15 47 21 07 82 4, Letter Series—odd -08 10 26 81 19 02 27 29 29 102 5. Letter Series—even 00 10 09 84 10 00 29 19 14 87 6, Unusual Details 25 30 07 28 21 15 41 25 29 62 7, Social Situations • ) 0 UU 38 12 38 23 17 43 02 04 62 8. Seeing Problems—I 00 11 15 07 83 17 14 13 28 87 9. Seeing Problems—II 16 04 09 06 86 03 00 09 17 82 10. Ideational Fluency—I-II 08 20 06 28 10 71 13 19 12 71 11, Ideational Fluency—ill-IV 15 2 1 18 08 07 62 15 34 29 72 12. Brick Uses—shifts 14 IS - 0 1 0 1 28 15 -08 74 16 74 13. Alternative Uses -02 22 15 0 1 26 08 21 64 27 67 14. Digit Symbol 54 01 06 36 16 34 10 10 -04 59 15. Vocabulary 10 79 24 09 07 06 22 17 20 82 16. Arithmetic 04 22 68 29 01 -01 04 15 14 64 Note: decimal points have been omitted. A-NSI, numerical facility F—DMU, ideational fluency B-CMU, verbal comprehension G—EMS, experiential evaluation C—CMS, general reasoning H--DMC, semantic spontaneous flexibility D—CSS, symbolic patterns R—residual E— EMI, sensitivity to problems TABLE 11 ROTATED FACTOR MATRIX, CONTROL GROUP 87 Variable A B C D E F G H R h^ 1. Numerical Operations 61 10 41 33 10 19 26 09 17 81 2. Verbal Comprehension 21 72 -03 33 16 14 10 25 01 79 3. Ship Destination -05 20 51 43 10 13 15 50 11 80 4, Letter Series—odd 14 35 25 79 26. 12 16 00 -17 97 5. Letter Series—even 06 13 26 85 25 16 23 09 13 98 6. Unusual Details 38 29 12 20 18 09 47 42 22 77 7, Social Situations 27 28 -01 13 23 22 37 20 14 47 8. Seeing Problems—I 25 14 15 02 73 20 09 37 14 84 9. Seeing Problems—II 18 02 13 19 66 02 15 63 -21 99 10. Ideational Fluency—I-II 20 22 14 07 05 69 05 20 26 70 11. Ideational Fluency—IH-IV 04 25 10 10 23 72 27 08 19 77 12. Brick Uses—shifts 08 09 00 01 09 51 10 58 49 75 13. Alternative Uses 10 08 07 25 24 16 15 61 29 65 140 Digit Symbol 63 10 05 24 06 25 15 14 16 60 15. Vocabulary 10 76 14 18 12 07 23 29 04 80 16. Arithmetic 00 uu 36 60 09 25 10 09 03 24 69 Note: decimal points have been omitted, A—NSlj numerical facility B--CMU, verbal comprehension C—CMS, general reasoning D—CSS, symbolic patterns E—EMI, sensitivity to problems F—DJwU, ideational fluency G—EMS, experiential evaluation H—DMC, semantic spontaneous flexibility R—residual than the convergent production of symbolic implica tions. The Digit Symbol test appears to require short-term memory, and Wechsler (99) points out a marked correlation between Digit Symbol and Memory Span for Digits. The appearance of Unusual Details on this factor, then, might reflect some memory content in Unusual Details for the control subjects. Factor B--CMU— verbal comprehension E 15. Vocabulary .79 x 2„ Verbal Comprehension .74 p 7 o Social Situations .38 EMS o 4 3 6. Unusual Details .30 EMS . 41 c 15. Vocabulary .76 o o i . d © Verbal Comprehension .72 CSS .33 n 16 < . Ari thmeti e . 3 6 CMS . 60 4 . Letter Series— -odd .35 CSS .79 It is not surprising that cognition of semantic units, or word meaning, should contribute to per formance on Social Situations, in which much reading is required, or Unusual Details, in which eight of the thirty-two items contain verbal cues. The replace ment of these two by Arithmetic and Letter Series in the control, group, and the separation of the odd and even items of Letter Series, are not readily expiainabl 89 Factor C— CMS— general reasoning E 16. Arithmetic .68 X 0 > • Ship Destination .56 EMS .47, P 1. Numerical Operations .35 NSI .45 C 16. Arithmetic . 60 CMU .36 0 O O Ship Destination . 51 Dr.! C . 50, n 1. Numerical Operations .41 NSI .61, Though studies with normal subjects would not predict it (10, 29, 33, 36, 39), cognition of semantic systems, or general reasoning, appears to enter into Numerical Operations for both groups of subjects in the present sample. Guilford (21) has pointed out that the factor composition of a test may change as its difficulty changes, and Zimmerman (110, 111) has sug gested that, as test items are made more complex, general reasoning is more likely to enter into their factor composition. It seems possible that the Numerical Operations items are relatively more complex and unfamiliar for schizophrenic than for normal sub jects, and therefore involve general reasoning for schizophrenics though not for normals. 90 Factor D— CSS— symbolic patterns E 5. Letter Series— even .84 X 4. Letter Series— odd .81 P 3. Ship Destination .41 CMS .56, EMS .47 7. Social Situations .38 EMS .43, CMU .38 14. Digit Symbol .36 NSI . 54, DMU .34 C 5. Letter Series— even .85 o 4. Letter Series— odd .79 CMU .35 n 3 . Ship Destination o 4 3 CMS DMC . 50 1. Numerical Operations e O O NSI .61, CMS .41 2. Verbal Comprehension . o 3 CMU .72 Looking first at the experimental group, vari ables 5, 4. 3, and 14 all require being able to grasp the nature of an unfamiliar system of symbols. Social Situations seems out of place, Its presence may signify a confounding of CSS variance with format variance; for in this study, variables 5, 4, 3, and 7, and they alone, require circling the right symbol to indicate the answer. Looking next at the control group, variables 5, 4, and 3 appear again. Numerical Operations also is a symbolic test, though its content is meant to be familiar• , with chronic schisophrenic subjects, the familiarity may be substantially reduced. The presence of Verbal Comprehension is not readily explainable. Factor E— EMI— sensitivity to problems E 9. Seeing Problems— II .86 x 8. Seeing Problems— I .83 C 8. Seeing Problems— I .73 DMC .37 n 9. Seeing Problems— II .66 DMC .63 The increased factorial complexity of Seeing Problems in the control group is of interest and will be discussed in connection with Factor H. .Evaluation of semantic implications, or sensitivity to problems, is the most clearly defined factor for the experimental group. Factor F— DMU— ideational fluency E 10. Ideational Fluency--I-II .71 x 11. Ideational Fluency---1II-IV.62 DMC .34 p 14. Digit Symbol .34 NSI .54, CSS .3s C 11. Ideational Fluency— III-IV.72 o 10. Ideational Fluency— I-II .69 n 12. Brick Uses (shifts) .54 DMC .58 Divergent production of semantic units, or ideational fluency, has not entered into Brick Uses (shifts) to any substantial degree with normal sub jects (32); with unlobotoraized schizophrenics, it does 92 here. The marked difference between the groups with respect to Brick Uses (shifts) will be dealt with in the discussion of Factor H. There is some previous support for relating Digit Symbol to fluency factors (10). This factor is the most clearly defined factor for the control group. Factor G— EMS— experiential evaluation E 3. Ship Destination .47 CMS .56, CSS .41 x 7. Social Situations .43 CMU .38, CSS .38 d •6. Unusual Details C 6. Unusual. Details .47 DMC .42, NSI .38 n 7. Social Situations .37 Evaluation of semantic systems, or experiential evaluation, appears in both groups much as expected, except for the presence? of Ship Destination in the experimental group. The lobotomized subjects are apparently more differentiated in the degree to which evaluation plays a part in their performance on that test, which is also a systems test, but primarily a cognitive one. Factor H— DMC— semantic spontaneous flexibility E 12. Brick Uses (shifts) .74 x 13. Alternative Uses .64 p 11. Ideationa1 Fluency— 1II-IV.34 0MU .S2 93 C 9. Seeing Problems— II .63 EMI .66 o 13. Alternative Uses 61 n 12. Brick Uses (shifts) .58 DMU .54 6. Unusual Details 3. Ship Destination .50 CMS .51, CSS .43 .42 EMS .47, NSI .38 8. Seeing Problems— I .37 EMI .73 Divergent production of semantic classes, also known as semantic spontaneous flexibility, appears in the experimental group much as expected. The presence of one of the Ideational. Fluency variables is under standable; a slight change in the response class on that test can increase the number of responses. ture from the normally predicted pattern. The two marker tests, Alternative Uses and Brick Uses (shifts), appear as predicted, except that a substantial portion of the variance of Brick Uses (shifts) has been lost to the ideational fluency factor, also a semantic divergent production factor. What is unexpected is that four other variables also appear. The phenomenon is orderly, since all four of the additional variables are semantic in content, and since flexibility of thinking can easily be seen as a contributor to their differentiation among subjects. In the control group there is a striking depar- Factor R— Residual No loadings greater than 0.29 in absolute value are found on this factor for either group. In the experimental group, six test variables were of complexities greater than one, having more than one significant factor loading. Complexity two: Numerical Operations, Unusual Details, and Ideational Fluency III-IV. Complexity three: Ship Destination, Social Situations, and Digit Symbol. In the control group, nine test variables were of complexities greater than one. Complexity two: Verbal Comprehension, Letter Series odd, Seeing Prob lems I and II, Brick Uses (shifts), and Arithmetic. Complexity three: Numerical Operations, Ship Destina tion, and Unusual Details. CHAPTER VII INTERGROUP COMPARISONS Hypothesis Testing The sixteen variables of the main test battery, and fourteen variables of the Wechsler-Bellevue Intel ligence Scale, Form I, are used to test the hypothesis that the control sub.jects have group mean scores that are not different from, or are lower than, those of the lobotomized sub.jects. Mean scores and variabilities for matched groups of 150 lobotomized and 150 control subjects on the main battery variables are presented and compared in Table 12. Ten of the sixteen variables show a significantly higher mean score for the control group. Though the intergroup differences are not great in absolute value, Numerical Operations, Verbal Comprehension, Ship Des tination, the Letter Series variables, Social Situa tions, the Seeing; Problems variables, and Alternative Uses differentiate the groups at a high level of significance. One Wechsler-Bellevue subtest, Arith metic , differentiates the groups at the 0.05 level. The other six variables show nonsignificantly higher scores for the control group. The first-stage com parison does not support the hypothesis. 95 TABLE 12 MEANS AND VARIABILITIES OF MAIN BATTERY VARIABLES 9 . Experimentals 150 Controls Ratios Variable Mean SD Mean SD F t I . Numerical Operations 41.89 12.51 47.49 15.57 1.55** 3.36 2. Verbal Comprehension 29.47 11.57 32.91 12.27 1.12 2.49 o « Ship Destination 14.28 5. 53 16.85 7.35 1.77** 3.35 4. Letter Series— odd 4.03 3.11 5.42 3.48 1.25 3.63 0 a Letter Series— even 4.31 D . 2 D 5.74 3.16 1.04 3.87 6 * Unusual Details ilo02 6.29 12.20 7.32 1.35* 1.49 m ( a Social Situations 11.15 3.22 12.27 3.85 1.43** 2.67 So Seeing Problems— I 4.75 4.22 6.80 5.19 1.51** 3.65 9« Seeing Problems— II o . 2 d 3.94 5.46 4.84 1.51** 4.20 10. Ideational Fluency— I-XI 13.38 6.36 13.87 6.42 1.02 0.66 11. Ideational Fluency— III-IY 17.05 8,37 18.67 8.67 1.07 1.64 12 . Brick Uses— shifts 2.63 3.OS 2.99 3.80 1.51** 0.87 X o o Alternative Uses 4.34 4.20 6.15 5.45 1.68** 3.16 14. Digit Symbol 29.09 9.45 29.15 10.52 1.23 0.05 15, Vocabulary 20.57 6.88 21.62 6.47 1.13 1.36 16. Arithmetic 5.48 2. 73 6.09 2.79 1.05 1.92 ^F-ratio has two-tailed test of significance; t-ratio has one-tailed test. One star indicates .05 level; two stars, ,01 level of significance or better. CD cn 97 Table 13 presents a similar comparison for matched groups of 69 lobotomized and 72 control white veterans. The variables are the eleven subtest scores and three IQ scores of the Wechsler-Bellevue Scale, Form I, The Digit Span and Similarities subtests, and the three IQ's, show significantly higher mean scores for the control subjects. The nine other subtests show nonsignificantly higher mean scores for the controls. The second-stage comparison does not support the hypothesis. Table 14 compares mean scores and variabilities of the ten factor-analytic tests (with parts recombined) for matched groups of 50 lobotomized and 50 control white veterans. Subjects have been eliminated in order to equate the mean and variability (cf. Table 1, page 59) of the IVechsler-Believue Full Scale IQ for the two groups. In spite of this matching on a correlated variable (11), Numerical Operations, Ship Destination, Letter Series, and Social Situations show significantly higher mean scores for the control group. The third- stage comparison does not support the hypothesis, which must be rejected. Factor-Defined versus Task-Defined Tests The foregoing comparisons suggest a superior ability of the factor-defined tests to differentiate TABLE 13 ■OIviPARISON OF MEANS AND VARIABILITIES OF WECHSLER-BELLEVUE SCALE. FORM I, 3UBTESTSC AND IQ's 69 b Experimentals 72 Controls*3 Rat a lOS Variable Mean SD Mean SD F t Information 1 o. 4 3 4.43 14.39 4.69 1.12 1 o 24 Comprehension 9.06 3. 94 9.60 3.86 1.04 0.82 Digit Span 9.35 1.88 10.08 2.10 1.25 2.15* Arithmetic 6.01 2.67 6.72 2 „ 83 1.13 1.51 Similarities 9.13 4.61 10.92 4.56 1.02 2.29* Vocabulary 21.62 6 o 64 22.28 6 o 64 1.00 Oo 58 Picture Arrangement 7.41 3.54 8.43 4.05 1.31 1.58 Picture Completion 8.90 3.19 9.72 3.14 1.03 1 o 53 Block Design 16.32 8.03 18.47 7.68 1.09 1.61 Object Assembly 16.93 4.88 17.51 4.08 1.44 0.76 Digit Symbol 28.91 9.52 30.07 10.02 1.11 0. 70 Verbal IQ 95o04 12.67 99.63 13.88 1.20 2.03* Performance IQ 96.90 15. 92 101o 49 13.77 1.34 1 o 82 * Full Scale IQ 95.70 13.49 100.46 14.22 1.11 2.03* ^F-ratio has two-tailed test of significance; t-ratio has one-tailed test. Star indicates o05 level of significance. '■'Matched groups of white veterans. wRaw scores. CD 00 c r c+ TABLE 14 COMPARISON OF MEANS ^ND VARIABILITIES OF FACTOR-ANALYTIC TESTS b b a Experimentals 50 Controls Ratios Test Mean SD Mean SD F t — © Numerical Operations 42,14 14.82 47.42 14. 39 1.06 1.79* V Verbal Comprehension oQ a 20 12.01 31.88 12.09 1.01 0.69 3 . Ship Destination 15.02 6.05 17.24 8.09 1.79* 1.75* Letter Series 8,48 5. 3 5 10. 58 6 . 3 2 1.40 1.78* e Unusual Details 13.06 6 o 33 12.76 7. 22 1.30 -0.22 ? o Social Situations 11.24 3.17 12.34 O a X 3 1.02 1.73* 8 - y » Seeing Problems 9.62 8,79 11.20 9.19 1.09 0.87 10-119 Ideational Fluency 31.92 13.11 31 . 24 13.75 1.10 -0.25 1 Q Brick Uses— shifts 3.18 0 . 1 / 2.88 4.33 1.87* -0.39 13. Alternative Uses 5,32 4.67 6.34 6.80 2.12** 0.86 F-ratio has two-tailed test of significance; t-ratio has one-tailed One star indicates o05 level; two stars, *01 level of significance etter. b Matched groups of whixe veterans, with equal group mean IQ's on the Wechsler-Bellevue Scale, Form I. to to 100 the lobotomized from the unlobotomized subjects. This performance is all the more impressive when one remem bers that the factor-defined tests were designed as group tests and have been little altered for this study, and that they tend to yield skewed score distributions, while the task-defined tests yield nearly normal dis tributions . Verbal Comprehension, Seeing Problems, and Alternative Uses discriminate between the groups before IQ matching, but not after. One may infer that they are more highly correlated with IQ than are Numerical Operations, Ship Destination, Letter Series, and Social Situations, whose ability to discriminate seems markedly independent of manipulation of the IQ. The greater articulation of our picture of intellect through the use of factor-analytic tests makes it possible to say not only that lobotomy results in a measurable impairment of intellect but also that the impairment appears in some areas of intellect though not in others. The results imply, for example, that lobotomized patients may be relatively impaired on one productive ability (flexibility) but not on another (fluency). The development of factor scores (95) anti their application in new studies (45) would help to clarify the pattern of factor-defined abilities 101 that are impaired by lobotomy in schizophrenia, and that may be impaired by other types of frontal lesions. Three factors, experiential evaluation (EMS), ideational fluency (DMU), and, particularly, flexibility (DMC), make different contributions to the factor com positions of tests for the two groups in this study. In support of Harvey's hypothesis (45), it may be shown with these results how a difference in a factor-defined ability may be obscured by a change in the factor com position of a task. Comparison of group mean scores (see Table 12, page 96) reveals the following indicative relationships. Ideational Fluency, representing the fluency factor primarily, does not differentiate the groups. It may be inferred from this that the groups do not differ in the factor-defined ability, ideational fluency. Alternative Uses, representing flexibility, does differentiate the groups, suggesting that the groups are different in the factor-defined ability, semantic spontaneous flexibility. Brick Uses (shifts), however, represents flexibility for the experimental group, but represents flexibility and fluency in about equal measure for the control group; it does not dif ferentiate the groups» If Brick Uses (shifts) were the only test employed, and if it were supposed to be a test of flexibility for both groups, then its failure 102 to differentiate would obscure the probable difference between the groups in the factor, semantic spontaneous flexibility. The foregoing illustrates a strength of the factor-analytic approach. It also serves as a warning that when tests are known to have a different factor composition for two groups, or when their factor com position is not known, caution must be exercised in interpreting raw score differences. Statements about intergroup differences, when factor loadings are dif ferent for the groups, should be based upon weighted scores that take the factor loadings into account. CHAPTER VIII INCIDENTAL FINDINGS AND FURTHER USE OF DATA Inspection of the correlation matrices, Tables 6 and 7, pages 75-78, reveals that the general level of intertest correlations is high in both groups. There are, in fact, only two nonsignificant correla tions, both in the experimental group. In another study (62) including Ship Destination, Social Situa tions, Verbal Comprehension, and Unusual Details, but using 204 normal adults as subjects, the average intercorrelation of these four tests, computed by the use of Fisher's z coefficient (23), is 0.24. In the present study, it is 0.55 for the experimental group and 0.47 for the control group. The difference between mean Fisher z coefficients for normals and lobotomized is significant beyond the 0.01 level; for normals and unlobotomized, it is significant beyond the 0.05 level; for lobotomized and unlobotomized, it is not signifi cant. It would appear that these schizophrenic subjects are individually different along some unmeas ured continuum that introduces a higher general level of intercorrelation among tests than is found with normals. 103 104 The degree of illness at the time of testing would seem the variable most likely to be responsible. Ratings of the degree of illness, by clinical personnel who knew the subjects, could not be used, owing to the geographical scatter of the subjects, the excessive number of raters needed, and the unobtainability of raters for some of the subjects. If, however, a check list of relevant items (61) could be applied to the 300 rather heterogeneous clinical records, as of the date of testing, useable indicators of degree of illness might be selected, New factor analyses, including two or more of these indicators, might clarify and simplify the factor structures that have been found. A look at the nontest rows of the intercorrela tion matrices reveals a number of provocative apparent relat.ionships* Weinstein and Teuber (102) report no relationship between level of schooling and decrement in general intelligence after brain injury in normals. Yet, in the present study, correlations of level of schooling with test variables appear different in the two groups; the degree of schooling seems to matter less to performance after lobotomy. In both groups, the males appear to do better ok the general reasoning tests. The paranoid and hebephrenic categories appear to have patterns of 105 correlation with tests that are distinct from each other and distinct in the two groups. In the lobotomized group, there are six significant negative correlations between test performance and being on a tranquilizer; in the control group there are none; possibly there is an interaction between lobotomy and tranquilizers that impairs performance. Smith (89) cites neurological studies of Yakolev, in which the length of interval from lobotomy to autopsy was positively correlated, with the amount of degenera tion of cerebral tissue. Smith (89, 90, 91) attributes the performance decrements of his topectomized subjects to neurological degeneration during an eight-year inter val. The postoperative interval in the present study ranges from five to fifteen years, and yet no significant correlation with test performance appears, except in the case of general reasoning, where the correlation seems to be an artifact mediated by the sex variable. Given the opportunity, the investigator proposes to carry this investigation beyond the original design; for there is a good deal of unanalyzed information remaining. A program of multiple correlations will be carried out, to determine which nontest variables merit facto.r-ana.lytic study. Then five additional factor- analyses will be done. 106 The first two will be identical to the present pair, except that selected nontest variables will be included. The third, fourth, and fifth will be based on data from the 141 white veterans. Intercorrelations will be recomputed for the twenty-four original vari ables, eight additional Wechsler-Bellevue subtests, and a group indicator. The degree-of-illness indicators also could be worked in at this point. The third analysis will include the main-battery test variables and the group indicator. The fourth will add selected nontest variables. The fifth will deal with all twenty-four test variables, and the group indicator. Both oblique and orthogonal rotations will be made, in order to extract as much information as possible. From the results of the fifth analysis it might be possible to develop a factor-score equation to maximize the predictive ability of the Wechsler-Bellevue Scale with respect to frontal lesions (45). CHAPTER IX DISCUSSION Within the limitations of the samples of tests and subjects, we have seen differences in the structure of intellect, and in the factor composition of tasks, for lobotomized and unlobotomized schizophrenics. The tasks are: beeing Problems, Ship Destination, Crick Uses (shifts), and Unusual Details. There are marked differences between the groups in the contributions made to test performance by evaluative and productive factors, especially the flexibility factor. Semantic spontaneous flexibility is a poorly differentiated ability for the control group, entering into many tests that have loadings on other factors; for the lobotomized group, it is well differentiated, entering into tests much as it would with normal subjects,, At the same time, we have seen that the per formance of the lobotomized subjects is inferior to that of the control subjects on most of the factor- analytic tests. Even after group matching on a test- correlated variable, the lobotomized subjects have lower mean scores on factor-analytic systems tests and a test that may reflect long-term memory. 107 108 Combining these observations, it appears that schizophrenic patients may have some kind of a problem with rigidity of thinking that is alleviated by a lobotomy, but at the cost of impairment of a number of factor-defined abilities. Just how a lobotomy could bring about a more normal contribution of flexibility to thinking is not easily explained. The impairment of the lobotomy group, in general, is not very great, in comparison with the control group. Possibly a relatively small additional impairment of some capacity, such as the ability to hold a set, or to maintain a systematic determining tendency, or to remain "programmed," or, as Pribram suggests (76), to hold on to or generate an intention, might modify the supposed rigidity of the schizophrenic without greatly aggravating the already existing schizophrenic impairment. It must be remembered that most of the impair ment has taken place before lobotomy (47, 50, 83, 84). One is left with the task of extrapolating from a small intergroup difference to the presumably greater difference that would accompany lobotomy in non- psychotic (97) or normal (55) groups. "It is still possible that long-term planning and initiative, creative work and capacity for radical readjustment 109 may be notably impaired by lesions of the frontal lobes" (46:23). The results tend to confirm the suggestion (45) that the attainment of the same end score on a test does not guarantee that the score has been attained in the same way. The mean scores of Brick Uses (shifts) and Unusual Details do not differentiate the subject groups, and yet their factor content is different for the two groups. Lobotomized and milobotomized subjects seem not to use the same abilities on these tests, though they achieve nearly the same scores. The more factorially complex the test, the less its score should reflect subtle realignments of factor composition. A score so factorially complex as the i V e c h s 1 e r - Be 1.1 e vu e It) (11) should respond only to gross, undifferentiated changes, in which a number of factor-defined abilities are markedly and simul taneously altered, in the same direction. Relatively univocal tests should respond to subtle changes in single factor-defined abilities. The present results appear to support this view. In the introductory section, the investigator said he would be satisfied with a study that would encourage and facilitate the application of multi- 110 factorial methods to problems in physiological, abnormal, and clinical psychology. It would seem that the results of this investigation should be encouraging and facilitating, for an old and familiar problem that stubbornly refused to yield to a task- defined approach has proved less impervious to a factor-defined approach, even with tests developed for quite a different population. Many of the difficulties attending the present investigation would be absent in studies of schizo phrenia or mental deficiency, or in studies making use of psychopharmacological variables. The present study comes too late for a thorough elucidation of the effects of psychosurgery, but it shows what might have been done, had suitable arrays of factor-defined tests been available. More important, it shows what can and should be done in the future. The surface has been scratched, but no more, in the development of factor-analytic tests for use with abnormal popu lations, or in clinical or physiologico-behavioral problems. But there can be little doubt whether such tests are needed. CHAPTER X SUMMARY Factor-analytic measurement has contributed little to physiologico-behavioral problems, in spite of Lashley's prediction, in 1941, that it would speed the coalescence of neurology and psychology,, At the same time, the literature on intellectual consequences of brain lesions has amply demonstrated the failure of "general intelligence" tests to analyze or, often, even to detect the effects of brain modifications such as those of psychosurgery. Work with animals is bringing the specification of unitary functional brain systems to a point where it outstrips psycho logical specification of unitary behavioral parameters. If research is to be fruitful, psychological variables must be os well conceived and specified as physiological variables. This investigation sought to improve measure ment technique in physiological, abnormal, and clinical psychology, by attacking an old but still unsolved problem, the measurement of intellectual effects of psychosurgery. Both traditional and contemporary methods of measurement were used, and their relative effectiveness was assessed. 111 112 A sixteen-variable factor-analytic test battery, relating to Guilford's structure-of-intellect model, was individually administered to 150 lobotomized schizophrenics and 150 schizophrenic controls, group- matched for age, schooling, duration of illness, sex, race, diagnostic subgroup, veteran status, hospital, and presence of a tranquilizer., Intermediate matched subgroups of 69 lobotomized, and 72 control, white veterans were tested with the Wechsler-Bellevue Intel ligence Scale, Form I. From these intermediate sub groups , a further selection was made, to equate group mean Wechsler-Bellevue IQ's for matched subgroups of 50 lobotomized and 50 control subjects. Nine factors were extracted from the inter correlations of the sixteen factor-analytic test variables, for each of the two main groups of 150 subjects. Graphic orthogonal rotations were made, to the criteria of simple structure, positive mani fold, and finally, psychological meaning„ Analytical rotations were made and compared« Eight interpretable factors and one residual appeared for each group. It was apparent that the two factor structures were not the same, the most striking dissimilarity involving the factor of semantic spontaneous flexibility, with regard to which the J obotomized group, more than the 113 control group, resembled normals. There were inter group differences in the factor composition of four of the testSo No suitable test of the statistical significance of the factor-structure differences was found to be available. Group mean test score comparisons were made for the large, intermediate, and small pairs of groups. The control subjects had significantly higher mean scores on most of the factor-analytic tests, on all throe Wechsler-Bellevue IQ ' s , and on throe Wechsler- Bellevue subtests. Four of the factor-analytic tests differentiated the groups that had been matched on Wechsler-Bellevue IQ. It was concluded that prefrontal lobotomy may alleviate an apparent abnormality of schizophrenics with regard to flexibility of thinking, but at the cost of impairment of a number of factor-defined intellectual abilities. The factor-analytic battery was found to be both more sensitive and more differ entiated than the general intelligence test in its discrimination between the groups, lending support to the contention that factor-analytic instruments and methods are indispensable for some phases of research in physiologica1, abnormal, and clinical psychology. If differential impairment of intellectual 114 abilities after lobotomy can be shown in subjects already much impaired by schizophrenia, through the use of factor-analytic tests, then the use of such tests with normal subjects and psychopharmacological variables should prove most fruitful. The analysis of schizophrenic thinking should also be possible, when factor-analytic tests are adapted or developed for that purpose. A program for further analysis of the data, and e o 11 e c t i on of additional data, was outlined. L J . O ( i O BIBLIOGRAPHY 1* Arkin, H. , and Colton, R. R. Tables for statis ticians . New York: Barnes and Noble, 1950. 2„ Burdock, E. I., Sutton, S., and Zubin, JD Personality and psychopathology. J. abnorra. soc. Psychol., 1958, 56, 18-30. 3. Caldwell, Anne E. Psychopharmaca, a bibliography of psychopharmacology, 1952-1957. U. S. Public Health Service, 1958. 4. Cat tell, R. B., and Baggaley, A0 R. The salient variable similarity index for factor matching. Brit. J. stat. Psychol., 1960, 13, 33-46. 5. Christensen, P. R., and Guilford, J. P. Ship destination test, form A. Beverly Hills, Calif o : Sheridan Supply Co nip any, 1955, 6. Christensen, P. R., and Guilford, J. P. Ideational fluency I, forni A. Beverly Hills, Calif. : Sheridan Supply Company, 1957. 7. Cobb, S. Personality as affected by lesions of the brain. In J. McV. Hunt (Ed.), Personality and the behavior disorders. New York: Ronald, 1944. Pp. 550-581. 8. Cohen, J. The factorial structure of the WAIS between early adulthood and old age. J. consult. Psychol., 1957, 21, 285-290. 9. Crown, S. An experimental study of psychological changes following prefrontal lobotomy. J. gen. Psychol., 1952, 47, 3-41. LQ. Davis, P. C. A factor analysis of the Vvechsler- Bellevue Intelligence Scale, Form I, in a matrix with reference variables. Doctoral dissertation, Univer0 of So. Calif., 1952. 11. Davis, P0 C. A factor analysis of the Wechsler- Bellevue Scale. Educ. psychol. Measmt., 1956, 16, 12,7-146. 116 117 12. Deese, J., and Morgan, C. T. Comparative and physiological psychology. Annu. Rev. Psychol., 1951, 2, 193-216. 13. Dixon, W. J., and Massey, F. J. Jr. Introduction to statistical analysis. New York: McGraw- Hill, 1951. 14. Dixon, W. J., and Massey, F. J. Jr. Introduction to statistical analysis. New York: McGraw- Hill, 1957. 15. Fisher, R. A., and Yates, F» Statistical tables for biological, agricultural, and medical research, fifth edition. London: Oliver and Boyd, 1957. 16. Freeman, W., and Watts, J. W. Psychosurgery: intelligence, emotion and social behavior following prefrontal lobotomy for mental disorders. Springfield, 111.: C. C. Thomas, 1942. 17. Fruchter, B. Introduction to factor analysis. New York: Van Nostrand, 1954. 18. Fulton, J. F. Frontal lobotomy and affective behavior, a neurophysiologicaf analysis. New York : Norton, 1951. 19,. Garner, Ann M. .Abnormalities of behavior. Avmu. dev. Psychol. , 1958, 9, 391-418. .... 20. Guertin, IV. II. , Frank, G. II., and Rabin, A. I. Research with the Wechsler-Bellevue Intelli gence Scale: 1950-1955. Psychol. Bull., I956, 53, 235-257. 21. Guilford, J. P. The difficulty of a test and its factor composition. Psychometrika, 1941, 6, 67-77. 22. Guilford, J. P. When not to factor analyze. Psychol. Bull., 1952, 49, 26-37. 23. 'killford, J. P. Fundamental statistics in psychology and edueat ion, third edition. blew York: McGraw-Hill, 1956. 118 24. Guilford, J. P. The structure of intellect. Psychol. Bull., 1956, 53, 267-293. 25. Guilford, J. P. A revised structure of intellect. Rep, psychol. Lab., No. 19. Los Angeles: Univer. of So. Calif., 1957. 26. Guilford, J. P. Creative abilities in the arts. Psychol. Rev., 1957, 64, 110-118. 27. Guilford, J. P. Personality. New York: McGraw- Hill, 1959. 28. Guilford, J. P. Factorial angles to psychology. Psychol. Rev., 1961, 68, 1-20. 29. Guilford, J. P., Berger, R. M., and Christensen, P. R. A factor-analytic study of planning. II. Administration of tests and analysis of results. Rep, psychol. Lab., No. 12. Los Angeles: Univer. of So. Calif., 1955. 30. Guilford, J. P., and Christensen, P. R. A factor- analytic study of verbal fluency. Rep. psychol. Lab., No. 17. Los Angeles: Univer. of So. Calif., 1956. 31. Guilford, J. P., Christensen, P. R., Wilson, R. C., and Merrifield, P. R. Alternate uses, Form A. Beverly Hills, Calif.: Sheridan Supply Company, 1959. 32. Guilford, J. P., Frick, J. \V. , Christensen, P. R. , and Merrifield, P. R. A factor-analytic study of flexibility in thinking. Rep, psychol. Lab., No. 18. Los Angeles: Univer. of So. Calif., 1957. 33. Guilford, J. P., Green, R. F., Christensen, P. R., Hertzka, A. F. , and Kettner, N. \V. A factor- analytic study of Navy reasoning tests with the Air Force Aircrew Classification Battery. Rep, psychol. Lab., No. 6. Los Angeles: Univer. of So. Calif., 1952. 119 34. Guilford, J . P., Hertzka, A. F., and Christensen, P. P. A factor-analytic study of evaluative abilities. II. Administration of tests and analysis of results. Rep, psychol. Lab., No. 9. Los Angeles: Univer. of So. Calif., 1953. 35. Guilford, J. P., kettner, N. W. , and Christensen, P. R. The relation of certain thinking factors to training criteria in the U. S. Coast Guard Academy. Rep. psychol. Lab., No. 13. Los Angeles: Univer. of So. Calif., 1955. 36. Guilford, J. P., Kettner, N. W., and Christensen, P. R. A factor-analytic investigation of the factor called general reasoning. Rep, psychol, Lab., No. 14. Los Angeles: Univer. of So. Calif., 1955. 37. Guilford, J. P., Kettner, N. \V., and Christensen, P. R. A factor-analytic study across the domains of reasoning, creativity, and evalua tion. II. Administration of tests and analysis of results. Rep, psychol. Lab., No. 16. Los Angeles: Univer. of So. Calif., 1956. 38c, Guilford, J„ P. , and Merrif ield, P. R. The structure of intellect model: its uses and implications. Rep, psychol. Lab., No. 24. Los Angeles: Univer. of So. Calif., 1960. 39. Guilford, J. P., Merrifield, P. R., Christensen, P. R., and Frick, J. W. An investigation of symbolic factors of cognition and convergent production. Rep, psychol. Lab., No. 23. Los Angeles: Univer. of So. Calif., i960. 40, Guilford, J. P., and Zimmerman, V,. S. The Guilford-Zimmerman aptitude survey. Beverly Hills, Calif.: Sheridan Supply Company, 1947, 41. Halstead, W. C. Brain and intelligence, a quantitative study of the frontal lobes. Chicago: Univer. of Chicago, 1947. 42, Hal stead, W„ C. , Carmichael, M. T., and .Bucy, P. C. Prefrontal lobotomy, a preliminary appraisal of the behavioral results. Amer. d . Ps.ychiat. , 1946, 103, 217-228. 120 43. Harlow, 11. 1’. Functional organization of the brain in relation to mentation and behavior. biology of mental health and disease. New York: Hoeber, 1952. Pp. 244-264. 44. llarman, Harry H. Modern factor analysis. Chicago: Univer. of Chicago, 1960. 45. Harvey, II. Cognitive changes in brain-injured children: a preliminary survey. doctoral dissertation, Univer. of So. Calif., 1950. 46. Hebb, Do 0. Man's frontal lobes: a critical review. A. M. A. Arch. Neurol. Psychiat., 1945, 54, 10-24. 47. Hebb, D. 0. Animal and physiological psychology. Anna. Rev. Psychol., 1950, 1, 173-188. 18. hertzka, A. F„, and Guilford, d. P. Logical reasoning, form A. Beverly Hills, Calif.: Sheridan Supply Company, 1955. 49. Hoel, P. G. Introduction to mathematical sta tistics . New York: Wiley, 1958. 50. Hoyt, Ruth, 1111 iott, H. , and Hebb, D. 0. The intelligence of schizophrenic patients follow ing lobotomy. Can. Dept. Veterans Affairs, Dull p Treatment Services, 195.1, 6, 553-557. 5.1. Kaiser, ii. IN The variraax criterion for analytical rotation in factor analysis, Psychometrika, 1958, 23, 187-200. 52. Kaiser, Ii. F. Relating factors between studies based upon different individuals, preliminary draft. Unpublished paper, I960. 53. King, H. E. Intellectual function. In F. A. Mettler (Ed.), Selective partial ablation of the frontal cortex. New York: Hoeber, 1949. 54. Klebanoff, S. G. Psychological changes in organic brain lesions and ablations. Psychol. Bull., 1945, 42, 585-623 0 121 55. Klebanoff, S. G. , Singer, J. L. , and Wilenslcy, H. Psychological consequences of brain lesions and ablations. Psychol. Bull., 1954, 51, 1-41. 56. Lashley, K. S. Brain mechanisms and intelligence: a quantitative study of in/juries to the brain. Chicago: Univer. of Chicago, 1929. 57. Lashley, Iv. S. Coalescence of neurology and psychology. Proc. Amer. phil. Soc., 1941, 84, 461-470. 58. Levinson, D., Meadow, A., Atwell, C., Robey, A. G.. and Beilis, E. The relation of frontal lobe surgery to intellectual and emotional function ing. In M. Greenblatt and H. C. Solomon (Eds.), Frontal lobes and schizophrenia. New York: Springer, 1953. Pp. 194-213. 39, Lewis, N. 1)„ C. , Landis, C. , and King, H. E. (Eds.). Studies in topectomy. New York: Grune and Stratton, 1956. 60. Lilly, J. C. Correlations between neurophysio- logical activity in the cortex and short-term behavior in the monkey. In H. F. Harlow and C. N. Wooisey (Eds.), Biological and bio chemical bases of behavior. Madison: Univer. of Wisconsin, 1958. Pp. 83-100. 61. Lorr, A!., Holsopple, J. Q., Jenkins, R» L. , and O'Connor, J. P. Factors of change in lohoto- mized chronic schizophrenic patients. J. consult. Psychol♦, 1955, 19, 39-43. 62. Marks, A., Guilford, J. P., and Merrifield, P. R. A study of military leadership in relation to selected intellectual factors. Rep, psychol. Lab., No. 21o Los Angeles: Univer. of So. Calif., 1959. 63. Markwell, E. D. Jr., Wheeler, W. M., and Kitzinger Helen. Changes in Wechsler-Bellevue test performance following prefrontal lobotomy. J. consult. Psychol., 1953, 17, 229-231. 66, Maxwell, A„ E. Statistical methods in factor 9 9 Vc 65. McReynolds, P., and Weide, Marian. The prediction and assessment of psychological changes follow ing prefrontal lobotomy. J. ment. Sci., 1959, 105, 971-978. 66. Merrifield, P. R. , Guilford, J. P., Christensen, P. R. , and Prick, J. W. A factor-analytic study of problem-solving abilities. Rep. psychol. Lab., No. 22. Los Angeles: Univer. of So. Calif., 1960. 67. Met tier, F. A. (Ed.). Select ive partial ablation of the frontal cortex, a correlative study of its effects on human psychotic subjects. New York: Hoeber, 1949. 68. Mettler, F. A. (Ed.). Psychosurgical problems. New York: Blakiston, 1952. 69o Miiner, Brenda. Intellectual function of the temporal lobes. Psychol. Bull., 1954, 51, 42-62. 70. Neff, V/. D. Physiological psychology. Anno. Rev. Psychol., 1953, 4, 255-272. 71. Nissen, H. W. , and Semmes, Josephine. Comparative and physiological psychology. Annu. Rev. Psychol., 1952, 3, 233-260. 72. Orbach, J. "Functions" of striate cortex and the problem of mass action. Psychol. Bull., 1959, 56, 271-292. 73, Perry, N. C., Kettner, N. W,, Hertzka, A0 F., and Bouvier, E. A. Estimating the tetrachoric correlation coefficient via a cosine pi table and correction graphs for non-median dichot- omization. Studies of Aptitudes of high-level Personnel: tech. Memorandum No. 2. Los Angeles: Univer. of So. Calif., 1953. 74. Pribram, K. H. Psychosurgery in midcentury. Surg. , Gyn. and Obstetr.', 1950, 91, 364-367. I n H o I1 ’. Bio logic a. Neocortical function in behavior, arlow and C. N. IVoolsey (Eds.), and biochemical bases of behavior. Madison: Univer. of Wisconsin, 1958. Pp. 151- 172. 123 76, Pribram, K. Si, A review of theory in physiological psychology. Annu. Rev. Psychol., 1960, 11, 1-40. 77. Pribram, K. H., and Fulton, J. F. An experimental critique of the effects of anterior cingulate ablations in monkey. Brain, 1954, 77, 34-44. 73. Pribram, K. H., and Kruger, L. Functions of the "olfactory brain." Ann. N. Y. Acad. Sci.., 1954, 58, 109-138. 79. Pribram, K. ii. , Mishkin, M. , Rosvold, ii. E. , and Kaplan, S. J. Effects on delayed response performance of lesions of dorsolateral and ventromedial frontal cortex of baboons. J. comp, physiol. Psychol., 1952, 45, 565-575. 80. Robinson, Mary F., and Freeman, W, Psychosurgery and the self. New York: Grime and Stratton, 1954. 81. Rosenzweig, I v i . ii. , Krech, D. , and Bennett, £. L. A search for relations between brain chemistry and behavior. Psychol. Bull., 1960, 57, 476- 492. 82. Ross, S., and Cole, J. 0. Psychopharmacology. Annu. iiev. Psychol. , 1960, 11, 415-438. 83. P.osvo 1 d, H. £. Physiological psychology. Amiu, iiev. Psychol. , 1959, 10, 415-454 . 84. Rosvold. H. E., and Mishkin, M. Evaluation of the effects of prefrontal lobotomy on intelligence. Can. J. Psychol., 1950, 4, 122-126. 85. kuja, D. H. Personality changes following pre frontal lobotomy in twenty-five schizophrenic patients. Doctoral dissertation, Univer. of So. Calif., 1953. 86. Scherer, 1, W. , Klett, C. J., and Winne, J. F. Psychological changes over a five year period following bilateral prefrontal lobotomy. J« consult. Psychol., 1957, 21, 25)1-295. 124 87. Scherer, I. vV. , Klett, C. J. , and Uinne, J. F. Changes in forty tests of functioning efficiency preoperation to five years after operation. Unpublished table, ADI 5180, 1957. 88. Semans, Catherine B. Use of the semantic differ ential with lobotomized psychotics. J. consult. Psychol., 1957, 21, 264. 89. Smith, A. Changes in Porteus maze scores of brain operated schizophrenics after an interval of eight years. Paper read at East. Psychol. Ass,, Philadelphia, April, 1958. 90. Smith, A., and Kinder, Elaine F. Effects of an eight-year interval on psychological measures of brain-operated (topectomy) schizophrenics. Amer. Psychologist, 1958, 13, 359. (Abstract) 91. Smith, A., and Kinder, Elaine F. Changes in psychological test performance of brain- operated schizophrenics after 8 years. Science, 19 5 9, 12 9, 149-150. 92. Stellar, E. Physiological psychology. Annu. Rev. Psychol., 1957, 8, 415-436. 93. Tcuber, ii. L. Physiological psychology. Annu. Key. Psychol., 1955, 6, 267-296. 94 a Thorndike, R. L. Individual differences, Annu. dev. Psychol., 1950, 1, 87-104. 95. Thurstone, L. L. Multiple factor analysis. Chicago: Univer. of Chicago, 1947. 96. Thurstone, L. L., and Thurstone, T. G. SKA primary mental abilities, intermediate, form AH. Chicago: Science Research Associates, 1947. 97. Tow, P0 M. Personality changes following frontal leucotomy. London: Oxford University Press, 1955, 98. Tucker, L. Rc A method for synthesis of factor analysis studies. Pers. Res. Sec. Rep. No. 984. Ar1ington: Armed Services Technica1 Information Agency, 1951. 125 99. 100. 101 . 102. . 1 \ > ' > < ■ > 104. 105. 106 » 1.07. 108. 109. Wechsler, D. The measurement of adult intelli- gence. Baltimore: Williams and Wilkins, 1944. Wechsler, D. Wechsler-Bellevue scale, for adolescents and adults, form I. New York: The Psychological Corporation, 1947. Wechsler, D. The measurement and appraisal of adult intelligence. Baltimore: Williams and Wilkins, 1958. Weinstein, S., and Teuber, U. L. The role of preinjury education and intelligence level in intellectual loss after brain injury. J. comp, physiol. Psychol., 1957, 50, 535-539. Wherry, P. .J. Determination of the specific components of maze ability for Tryon1s bright and dull rats by means of factor analysis. J. comp. Psychol., 1941, 32, 237-252. White, R. W. The abnormal personality. New York: Ronald, 1956. White, R. Wr. Abnormalities of behavior. Annu. Rev. Psychol., 1959, 10, 265-286. Uidemaii, Ii. Psychological investigations in the lobotomy study. In A. tliller, Lobotomy, a clinical study. Ontario: Ontario Dept, of Health, 1954. U'oifle, D. Factor analysis to 1940. Psychometric Monogr., 1940, 3. Woolsey, C. N. Organization of somatic sensory and motor areas of the cerebral cortex. In H. F. Harlow and C. N. Woolsey (Eds.), Biological and biochemical bases of behavior Madison: Univer. of Wisconsin, 1958. Pp. 63 81. Zimmerman, W. S. A simple graphical method for orthogonal rotation of axes. Psychqmetrika, 1946, II, 5I-55. 110. 111. 112. 12 6 Zimmerman, \V. S. The influence of item complexity upon the factor composition of a spatial visualization test. Educ. psychol. Measmt., 1954, 14, 106-119. Zimmerman, VV. S. Hypotheses concerning the nature of the spatial factors. Educ. psychol. Measmt., 1954, 14, 396-400. Zubin, J. Abnormalities of behavior. Annu. Rev. Psychol., 1952, 3, 261-282.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

A Factor-Analytic Study Of Military Leadership

PDF

Creativity In Children: A Study Of The Relationship Between Temperament Factors And Aptitude Factors Involved In The Creative Ability Of Seventh Grade Children With Suggestions For A Theory Of C...

PDF

A Study Of The Relationships Between Factors Found In Cattell'S Sixteen Personality Factor Questionnaire And Factors Found In The Guilford Personality Inventories

PDF

On The Relationship Between Anxiety And Aggression In Nine-Year-Old Boys

PDF

Intellectual And Cognitive Factors In The Production Of Psychological Stress Reactions

PDF

Behavioral (Social) Intelligence: A Factor Analysis

PDF

A Factor Analysis Of The Semantic-Evaluation Abilities

PDF

Human Performance As A Function Of The Joint Effects Of Drive And Incentive Motivation

PDF

A Factor-Analytic Study Of Problem-Solving Abilities

PDF

The Development Of Attitude Scales To Predict Accident Repeater And Moving Violator Drivers

PDF

Non-Specific Treatment Factors And Deconditioning In Fear Reduction

PDF

The Effects Of Prior Part-Experiences On Visual Form Perception In The Albino Rat

PDF

An Application Of Adaptation Level Theory To The Response Bias Of Falsification

PDF

An Analysis Of The Selection Criteria For Assignment Of Students To Advanced Placement Classes In The Los Angeles Unified School District

PDF

A Factor Analysis Of The Symbolic-Evaluation Abilities

PDF

A Factor Analysis Of The Figural-Evaluation Abilities

PDF

The Effect Of Motor Ability Loss On Cognition And Emotion

PDF

Sex-Role Preferences Of Early Adolescents In Relation To Adjustment

PDF

A Critical Analysis Of The Philosophical Fragments Of Epicharmus

PDF

Factorial Stability As A Function Of Analytical Rotational Method, Type Of Simple Structure, And Size Of Sample

Asset Metadata

Creator De Mille, Richard (author)

Core Title Intellect After Lobotomy In Schizophrenia: A Factor-Analytic Study

Degree Doctor of Philosophy

Degree Program Psychology

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest,Psychology, clinical

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Harvey, Herman (committee chair), Guilford, Joy P. (committee member), Werkmeister, William H. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-243461

Unique identifier UC11358089

Identifier 6103816.pdf (filename),usctheses-c18-243461 (legacy record id)

Legacy Identifier 6103816.pdf

Dmrecord 243461

Document Type Dissertation

Rights De Mille, Richard

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA