EFFECT OF PERFORMANCE FEEDBACK ON DEPRESSED AND NONDEPRESSED PSYCHIATRIC PATIENTS by SUSAN KAY MCBRIDE B.A., Kent State University, 1967 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in the Department of Psychology We accept this thesis as conforming to the required standard 'THE UNIVERSITY OF BRITISH COLUMBIA February, 1970 In present ing th i s thes is in pa r t i a l f u l f i lmen t o f the requirements fo r an advanced degree at the Un ivers i ty of B r i t i s h Columbia, I agree that the L ibrary sha l l make it f r ee l y ava i l ab le for reference and study. I fu r ther agree tha permission for extensive copying o f th i s thes is fo r scho la r l y purposes may be granted by the Head of my Department or by h is representat ives . It is understood that copying or pub l i c a t i on , of th is thes is fo r f inanc ia l gain sha l l not be allowed without my wr i t ten permiss ion. Department The Univers i ty of B r i t i s h Columbia Vancouver 8, Canada Date ABSTRACT The present study investigated the effects of success and failure performance feedback on subjects' (a) prediction of their performance on a future task, (b) estimation of parental and stranger predictions of their performance, (c) post-task evaluation, and (d) conformity behavior. High and Low Depressed psychiatric patients, selected on the basis of the Beck Depression Inventory (DI) (Beck, 1961, 1967), were given an experimentally induced success, failure, or neutral experience on a task of time estimation. They were then asked to predict their performance on a task of matching geometric figures. The conformity situation involved subjects' estimation of line lengths, after being given an erroneous hint as to the length of each line . The only measure significantly related to depression level was the magnitude of conformity responses, on which Low Depressed patients gave a greater number of inches of error i n the hinted direction than High Depressed patients. There was no difference between these two groups on the frequency of conformity responses. It was suggested that the lack of hypothesized differences between High and Low Depressed subjects may have been due to ineffectiveness of the DI in differentiating between depression levels, or due to limitations of Beck's theory in predicting the behavior of depressed patients in a nonsocial experimental situation. The lack of differences between the three experimental groups was probably due to aspects of the experimental situation which decreased the effectiveness of the performance feedback. i i TABLE OF CONTENTS Page Abstract i Table of Contents i i L i s t of Tables i i i Chapter One - History of the Problem and Background of the Present Study 1 Chapter Two - Method 12 Chapter Three - Results 29 Chapter Four - Discussion 41 Chapter Five - Summary 47 Footnotes 51 References 52 Appendix A - The Beck Depression Inventory with the Score Value for Each Alternative 55 Appendix B - Stimulus Cards Used in the Matching Task 59 Appendix C - The Six Line Stimulus Cards Used for the Conformity Task 91 Appendix D - The Five-Point Continuum Stimulus Card Used for the Post-Task Evaluation 99 Appendix E - Comparison of the Means of the 42 Subjects Who Were New Intake Patients and the Total 48 Subjects 101 i i i LIST OF TABLES Page Table 1. Age of Subjects In Each Group 14 Table 2. Marital Status of Subjects in Each Group 14 Table 3. Educational Level of Subjects in Each Group . . . . 15 Table 4. Occupation, Classified by Level of Function, of Subjects i n Each Group 16 Table 5. Parental Educational Level of Subjects in Each Group 17 Table 6. Diagnostic Classification of Subjects in Each Group 18 Table 7. Age at Fi r s t Psychiatric Treatment, and the Number of Years Ago Treatment was F i r s t Received for Each Group 20 Table 8. Frequency of Previous Hospitalization, Outpatient Care, and Shock Treatment for Each Group 21 Table 9. Means of the Post-Task Evaluation 30 Table 10. Mean Inches of Error i n the Hinted and Nonhinted Direction for a l l Conformity Items 31 Table 11. Analysis of Variance of the Magnitude of the Conformity Response 32 Table 12. Mean Inches of Error i n the Hinted Direction on A l l Conformity Items 32 Table 13. Frequency of Each Type of Response for Both Depression Levels . 33 Table 14. The Frequencies and Percentages of Both Types of Nonaccurate Responses for Each Item 34 iv Page Table 15. Analysis of Variance of the Four Performance Predictions . . . . . . 3 6 Table 16. Means for the Four Performance Predictions for Each Group 37 Table 17. Means for the Four Performance Predictions for A l l Subjects 3 8 Table 18. Mean Performance Predictions for the Significant Three Factor Interaction 3 8 Table 19. Mean Errors Made on the Matching Task 40 CHAPTER ONE HISTORY OF THE PROBLEM AND BACKGROUND OF THE PRESENT STUDY Based on the findings of recent research and c l i n i c a l experience, Beck (1967) has advanced an etiological theory of depression to explain what he terms the "primary triad in depression" (p. 255), The present study investigated some of the hypotheses derived from this descriptive triad and theory. According to Beck, the characteristics of depressives are mani-festations of the dominance of three cognitive patterns: viewing one's self, one's experiences, and one's future in a negative way. These cognitive patterns become dominant i n the depressive state and, as screening and interpretive mechanisms, they "divert the thinking into specific channels that deviate from rea l i t y " (Beck, 1967, p. 273), forming the basis of the cognitive, affective, motivational, and physical characteristics of depression as observed by clinicians. Aside from c l i n i c a l observations describing depressed patients in ways consistent with Beck's triad of cognitive patterns, there is also some experimental evidence that depressed patients are highly self--c r i t i c a l and defeatist, and interpret experiences and foresee their future in negative ways. Beck and Stein (I960), using a 5-point scale, compared the self-concept score arid the self-acceptance Score of depressed and nondepressed psychiatric patients. By correlating these two scores with the Beck Depression Inventory (DI) score (Beck, 1961, 1967) for each patient, they found significant negative correlations between both the DI score and the self-concept score, and the DI score and the self-2 acceptance score, indicating that the depressed patients had a more negative self-concept and displayed more self-rejection than did the nondepressed patients. Laxer (1964), using the Semantic Differential (Osgood, Suci, & Tannenbaum, 1957), found that depressed patients, when compared to paranoid patients, showed a lower self-concept on admission to hospital but shifted to a higher self-concept when discharged. An obvious problem inherent in the above studies i s that of response style. Neuringer (1963), Zax, Gardiner, and Lowy (1964), and Arthur (1966) have a l l found that psychiatric patients, when responding to tests such as the semantic differential, select significantly more extreme responses (1 and 7, or 1 and 5) than do nonpsychiatric subjects. Although the above studies carried out comparisons within a psychiatric population, i t is possible that the differences found were confounded by the existence of a differential response bias among psychiatric groups. Beck (1967) cites an unpublished study (Loeb, Beck, Diggory, & T u t h i l l , 1966) which employed a card sorting task to investigate the difference in performance level and self-confidence between 20 depressed and 20 nondepressed psychiatric patients, as determined by their DI scores. While the performance levels did not differ significantly, the depressed patients were significantly more pessimistic about their likelihood of succeeding and had a tendency to underrate the quality of their performance. Although there have been few experimental studies of this nature, the evidence thus far supports Beck's description of psychotic depression. The etiological theory Beck has advanced to explain the development and progressive dominance of these cognitive patterns and characteristics 3 in depression is primarily environmental. Beck states that an individual's concepts about himself and his world are derived from his experiences and develop early in l i f e , through interaction and identification with "key figures such as parents, siblings, and friends." Further,"onee a particular attitude or concept has been formed, i t can Influence subsequent judgments and become more firmly set" (Beck, 1967, p. 275). For example, in the case of depression, the child may experience one or a series of failure situations in which he receives blame for the failure and through identification feels self-blame. According to Beck's conception, these experiences are assimilated into the developing self-concept, which is employed in the child's inter-pretation of future experiences. Thus, a cycle of self-judged failure arid self-blame is established and the individual becomes sensitized to these particular situations (e.g., failure) which may evoke undue stress, and, at some later point, lead to severe depression. Beck's theory thus states that the pre-depressive experiences certain traumatic situations of a specific nature in his developmental years. These traumatic experiences are assimilated into the child's self-concept and cognitive network and become screening and interpretive agents for future experiences. Thus, Beck describes a circular feedback and interpretive system whereby certain cognitive patterns are developed by experience and strengthened by their own function in interpreting future experiences. In addition, this model of early traumatic experiences implies that the pre-depressive individual w i l l have areas of "specific vulnerability" (Beck, 1967, p. 277) which, together with certain environmental conditions, can cause overwhelming stress and 4 severe depression at some later point in his l i f e . A number of investigators have suggested the existence of a particular family structure and environment which, they hypothesize, is conducive to the development of these cognitive patterns and depressive symptoms. Cohen, Baker, Cohens Fromm-Reichmann, and Weigert (1954) investigated the background of 12 manic-depressives and concluded that in a l l 12 cases the patient's parents were highly concerned with status and social prestige, and were involved in a struggle to improve their position and acceptability in the community. Symbols of prestige were highly valued, and pressure was placed on the child to conform to social standards and to compete in the achievement of the. necessary symbols that would raise the family's status. Similar findings were reported by Gibson (1957), when he compared the l i f e histories of depressed and schizophrenic patients. He also found that his depressed patients came from homes in which the parents were greatly concerned about social approval and were involved in a struggle for prestige, accompanied by envy and competitiveness. Both Cohen et a l . (1954) and Gibson suggested that the patient was often selected by the parents, especially the mother, to be the child who should most try to f u l f i l l their needs for prestige. Additional evidence that parental pressure to conform to accepted standards i s common in the l i f e histories of depressive patients comes from an early study by Wilson (1951), again using the case records of 12 manic-depressive patients to arrive at his conclusions. Katkin, Sasmor, and Tan (1966) investigated the hypothesis that depressed patients would show more conformity behavior than nondepressed 5 patients in what they termed an Asch-type conformity situation. The conformity situation in this study did not involve the presence of other people, but the presence of other people presenting opinions on a tape recorder. After listening to the recorded opinions, patients were asked to indicate their own opinion on a continuum scale. These responses were then compared with the patients' responses prior to listening to the recorder. As expected, the responses of depressed patients showed more change in the direction of the recorded opinions than did the responses of nondepressed patients. Becker (1960), using McClelland and his associates' (McClelland, Atkinson, Clark, & Lowell, 1953; DeCharms, Morrison, Reitman, & McClelland, 1955) distinction between need Achievers and value Achievers to describe the manic-depressive and his background, suggested that manic-depressive patients resemble the high value Achievers who value achievement for the prestige involved, and fear failure for similar reasons. As to the origin of the high value achievement orientation, DeCharms jat a l . (1955) suggested that high value Achievers, similar to the description of depressives given by Cohen et a l . (1954), suffered from "authoritarian pressure from their parents to be ambitious" (p. 419). Thus, Becker maintains, .manic-depressives and high value Achievers are similar in their use of conformity, competitiveness, and achievement to win social approval, and their excessive fear of f a i l i n g to achieve this desired approval. In a study designed to investigate these hypothesized similarities, Becker (1960) administered the Value Achievement Scale (DeCharms et a l . , 195J) j the California Fascism Scale (F Scale) (Adorno, Frenkel-Brunswik, Levinson, & Sanford, 1950), and the Traditional Family Ideology Scale. 6 (TFI Scale) (Levinson & Huffman, 1955) to 24 remitted manic-depressives and 30 nonpsychiatric controls. The depressive subjects were selected on the basis of c l i n i c a l judgments and results of the Hildreth Mood Scale (Hildreth, 1946). On the basis of these measures, and two additional tasks, one with task-oriented instructions and one with achievement-oriented instructions, Becker concluded that his depressed patients "performed like high value Achievers with respect to achievement striving, attitudes toward authority, and conventional beliefs." He further concluded that for these subjects, "achievement and conformity behavior seem to have the goal of gaining the approval of others rather than of satisfying internalized standards" (p. 337). In a similar study, Spielberger, Parker, and Becker (1963) administered the F Scale, TFI Scale, Value Achievement Scale, and the Need Achievement Scale (McClelland et a l . , 1953) to 30 remitted manic-depressives and 30 non-psychiatric controls. Their conclusions were in agreement with Becker's earlier findings. An important limitation of Becker's (1960) and of Spielberger's et a l . (1963) investigations i s the failure to use hospitalized patients as controls. The importance of this i s demonstrated by another study by Becker, Spielberger, and Parker (1963) in which they administered some of the same scales used in the other studies to a group of neurotic depressives, schizophrenics, manic-depressives, and normal controls. They found no significant differences in F Scale scores or Value Achievement scores within the psychiatric groups, but did find significant differences between the psychiatric groups and the normal controls. Similarly, the study by Katkin e_t a l . (1966), which used hospitalized patients as a control, failed to find significant differences on the F Scale, Value Achievement Scale, or TFI Scale, further questioning the results of the earlier two studies. 7" The above studies suggest a picture of the depressive as coming from a family where great value is placed on social approval and status, with high parental expectations and pressures to conform and achieve the desired prestige. However, conclusions drawn from such studies as these are limited by methodological problems such as inadequate controls for the effects of hospitalization discussed above. In addition, the use of case records as the basic data by several investigators assumes consistency and accuracy on the part of hospital personnel, as well as allowing for bias in the interpretation of ^jtie histories. Becker's (1960) and Spielberger et al.'s (1963) results have come under specific attack by results of the study by Becker et a l . (1963), which found age and social class to significantly affect scores on measures used by the two earlier studies, pointing out the need for better control of such variables. The present study was designed to investigate some of the hypotheses derived from these earlier studies and Beck's theory, within a framework which would control for some of these problems. One hypothesis investi-gated was that there would be a differential response between depressed and nondepressed subjects due to the experiencing of success and failure. Loeb, Feshback, Beck, and Wolf (1964) attempted to test this position using the specific hypotheses that a success, as compared to a failure experience, would result "in positive affect, in self-confidence, and i n attribution of happiness to others," and that the depressed patients would be more affected by the experimentally induced success and failure thar would nondepressed patients" (p. 610). Their design, however, confounded the effects of success and failure x^ith that of competition. 8 Specifically, they asked 20 depressed male patients and 22 non-depressed male patients, selected on the basis of their DI scores, to rate, on an 11-point scale, the mood of 14 individuals, presented by photographs, as well as the patients' own mood. Following this, a l l patients were seated in a conference room and given a l i s t of 20 i n -complete words, which they were asked to complete in three minutes. Loeb et a l . manipulated the performance of the patients by varying the number of blanks they had to complete. Patients assigned to the success group had fewer blanks to complete in the allotted time and thus performed at a higher level. The performance scores of a l l patients were listed on a blackboard alongside each patient's name before they l e f t the room. Patients were then again asked to rate their own mood and the mood of the 14 photographed individuals, using the same scale. In addition, each patient was asked (a) to indicate the number of words he thought he could write in three minutes, and (b) i f he would volunteer to compete with other men on a slightly more d i f f i c u l t task at a later date. The responses to these last two questions were used as an indication of the patient's level of confidence after the task was completed. The results indicated that the success and failure experiences did affect the patient's mood ratings and that the changes for the depressed patients were about twice as great as were the changes for the nondepressed patients. Results of the ratings of the moods of the 14 photographs indicated no significant difference between depressed and nondepressed patients before or after the word completion task, although for both depressed and nondepressed patients in the success group the post-task ratings of the "happy" faces were significantly happier than the pre-task ratings. Loeb et al.'s 9 third variable, level of confidence, is the one of most concern to the present discussion. These results indicated that volunteering for a future task is affected by level of performance, but unrelated to degree of depression. However, analysis of the word estimates indicated that both level of performance and degree of depression affect this measure. Both superior performance groups gave higher mean estimates than the inferior performance groups, but the depressed superior performance group gave the highest estimate. The depressed inferior performance group gave the lowest estimate, although this was not signi-ficantly different from the estimate of the nondepressed inferior performance group. These results appear to indicate that the success and failure experiences did have a differential effect on the depressed and non-depressed subjects, with the depressed subjects being more affected, when measured by their estimated future performance. However, such a conclusion would be erroneous due to the confounding of success and failure experiences with competitiveness. Success (failure) in Loeb et al_.'s design does not lead to "I did well (poorly) on this task," but rather to "I did better than (worse than) others in this task." The earlier discussion of psychotic depression indicated that social approval and competition may be highly important aspects of the depressive's background and vulnerabilities, and thus the information that he performed better than others may have a different effect on the depressive than the information that he did well on the task. For this reason, the present study was designed to eliminate the confounding variable of competition and social comparison to enable the 10 investigator to study the effects of success and failure on: (a) depressive subject's prediction of his future performance on a specific task; (b) depressive subject's estimate of his mother's, father's, and a stranger's prediction of his future performance on the task; (c) depressive subject's estimate of the level of his performance on the task after the task had been completed; (d) the amount of conformity behavior displayed. In addition, the design allowed comparisons to be made between depressed and nondepressed patients on a l l these measures but without the experimentally induced success or failure experiences. A l l comparisons were carried out between High Depressed and Low Depressed psychiatric patients, selected on the basis of their DI scores. The basic design involved an experimentally induced success, failure, or neutral experience through the administration of performance feedback on a task of time estimation. Following this a l l patients were asked to give predictions for their performance on a task of matching geometric figures. The conformity situation involved subjects' estimation of the lengths of six lines, presented individually on cards. Patients were given an incorrect hint as to the actual length of each line. The f i f t h measure employed was the subject's post-task evaluation of his performance on the matching task. Beck's theory and the results of earlier studies suggested the depressive suffered from excessive parental expectation to achieve and succeed, with pressure placed on conforming in order to win social status and approval. Depressives were described as highly s e l f - c r i t i c a l , defeatist, habitually underestimating their a b i l i t i e s and performance, and specifically vulnerable to success and failure experiences. Thus, i t 11 was expected that High Depressed subjects, compared to Low Depressed subjects, would: (a) Give a lower prediction of their future performance on a specific task. (b) Give a lower parental and stranger prediction of their future performance. (c) Give a lower post-task evaluation of their performance. (d) Show more conformity behavior. And f i n a l l y , that: (e) Success would raise the predictions and post-task evaluation, and lower the amount of conformity behavior for a l l subjects, failure having the opposite effect, but High Depressed subjects would be more affected by the performance feedback than Low Depressed subjects. Maternal and paternal estimated predictions were expected to be similarly affected; however, since Cohen et_ al_. (1954) and Gibson (1957) had suggested that the mother was more instrumental in the early emphasis on achievement, these predictions were made separately. The stranger's estimated prediction of performance was included as a check on the possibility that the parental estimated predictions would not be specific to the parents, but rather reflect a general tendency to regard a l l people as degrading the depressive's a b i l i t y . Thus, there were two kinds of comparisons carried out in the present design: comparisons which explored differences between depressed and non-depressed patients on the variables mentioned, and comparisons which investigated the differential effect of success and failure on depressed and nondepressed patients. CHAPTER TWO METHOD Subjects 1 A l l subjects were patients at Vancouver General Hospital Adult Outpatient Department of Psychiatry. Vancouver General Hospital i s a teaching hospital, a f f i l i a t e d with the University of British Columbia. The department is staffed by a chief psychiatrist, several part time psychiatrists, a varying number of residents in training, two community health nurses, and one M.A. Psychologist. During the year 1968, 497 patients were referred to the department for psychiatric treatment, and 415 of these had had no previous contact with the department. To be eligible for treatment from the department, a patient must either be receiving some form of social assistance, or have income within the following limits: Single person $150.00 per month Two person family $170.00 per month Four person family $225.00 per month Six person family $275.00 per month. According to the Canadian St a t i s t i c a l Review (1968), this "single person" income is within approximately the lower 39% of Canadian incomes in 1968, and this "two person family" income is within approximately the lower 21%. In addition, no patients are treated at the department i f they have any form of medical insurance which would cover the treatment needed. The number of new patients referred to the department in any month varies, with fewer being treated during the summer months. Patients for the present study were seen between March 20, 1969 and October 7, 1969. The study was composed of six groups of subjects: two levels of 13 depression within each of three experimental conditions. There were four males and four females in each group, making a total of 48 subjects, 24 Low Depressed and 24 High Depressed. It was originally intended that a l l patients would be seen during their f i r s t v i s i t to the department. However, during the data collection i t was discovered that the number of new Low Depressed patients coming to the department was not as great as had been anticipated. For this reason, i t was necessary to include as subjects six Low Depressed patients who were not new intake patients at the hospital. Of these six, four were after-care patients, having been hospitalized at either Vancouver General Hospital or the Provincial Mental Hospital; and two, although not after-care patients, had been admitted to the department prior to the start of this study. These six subjects were in the following groups: two females in the Low Depressed, failure group; one male and one female in the Low Depressed, success group; and one male and one female in the Low Depressed, neutral group. Possible problems created by including these six patients w i l l be discussed later. Although a l l patients were randomly assigned to the three experi-mental groups, with no attempt made to match the groups on any variables other than sex, some demographic data and information concerning previous psychiatric treatment were recorded and are presented in Tables 1 to 8. As Table 1 indicates, the age range of the 48 subjects was between 18 and 62, with the majority being between 20 and 40. Although i t was originally intended to have a more limited age range, this was expanded due to the problem of obtaining subjects. Table 2 indicates that the majority of a l l subjects were single. 14 Table 1 Age of Subjects in Each Group Group 18-19 20-29 30-39 40-49 50-59 60-62 High Depressed Success 3 4 1 0 0 0 Neutral 0 4 4 0 0 0 Failure 0 3 3 1 0 1 Mean age: 28.91 Median age: 27 years years Low Depressed Success 0 4 2 0 1 1 Neutral 0 3 2 2 1 0 Failure 1 4 1 0 2 0 Mean age: 34.58 Median age: 31 i years years Table 2 Marital Status of Subjects in Each Group Group Married Single Separated Divorced Widowed High Depressed Success 1 7 0 0 0 Neutral 2 5 1 0 0 Failure 0 4 1 2 1 Low Depressed Success 3 3 1 1 0 Neutral 2 6 0 0 0 Failure 2 5 0 1 0 15 Of the 10 who were married, 8 were females. A l l subjects were required to have at least a sixth grade education. However, as shown in Table 3, 47 of the 48 had at least an eighth grade education, with the majority being within grades 8 through 11. Table 3 Educational Level of Subjects in Each Group Group Grades Some University University 6-7 8-11 12 Education Graduate High Depressed Success 0 6 2 0 0 Neutral 0 6 1 1 0 Failure 1 6 0 1 0 Low Depressed Success 0 5 2 0 1 Neutral 0 7 1 0 0 Failure 0 4 2 2 0 The occupation of each patient was recorded and classified according to Roe's (1956) system of classification by level of function, which is based on the "degrees of responsibility, capacity, and s k i l l " (p. 149). These data, and the six levels of this classification system, are presented in Table 4. For those subjects who were unemployed at the time of the study, their most recent occupation was recorded. As Table 4 indicates, the majority of patients reported occupations at levels 5 and 6, which would involve a lesser degree of responsibility, capacity, and s k i l l than the higher level occupations. 16 Table 4 Occupation, Classified by Level of Function, of Subjects in Each Group Group Levels 1 2 3 4 5 6 High Depressed Success 0 0 0 2 1 5 Neutral 0 0 1 2 3 2 Failure 0 0 0 3 0 5 Low Depressed Success 0 1 0 2 2 3 Neutral 0 0 0 1 4 3 Failure 0 0 1 2 2 3 Levels: 1. Professional and Managerial I (Independent Responsibility) 2. Professional and Managerial II (Medium-Level Responsibility) 3. Semi-Professional and Small Business 4. Skilled 5. Semi-Skilled 6. Unskilled In addition to this information, the educational level of the parents of each subject, as reported by the subject, was also recorded. This information is presented i n Table 5. As noted in the table, information concerning the parental educational level of two patients was impossible to ascertain either from hospital records or from the patient. As shown in the table, the educational level of both mother and father, for the majority of subjects, was grade 11 or below. 17 Table 5 Parental Educational Level of Subjects in Each Group Group Grades Some University University 1-8 9-11 12 Education Graduate Father's Education; High Depressed Success 2 . 3 2 0 1 Neutral'-' 2 3 1 0 1 Failure 4 2 1 1 0 Father's Education; Low Depressed Success* 3 4 0 0 0 Neutral 5 3 0 0 0 Failure 3 3 0 0 2 Mother's Education; High Depressed Success 2 2 3 1 0 Neutral* 3 3 1 0 0 Failure 3 4 1 0 0 Mother's Education; Low Depressed Success* 4 3 0 0 0 Neutral 3 3 2 0 0 Failure 1 4 2 0 1 *This information unknown for one subject. Diagnostic classification of each patient was obtained through hospital records and is presented in Table 6. Of the five subjects lis t e d as requiring counseling for specific problems, three required occupational counseling and two marital counseling. Only one of these five had any previous psychiatric treatment and this consisted of one outpatient 18 department contact within a year prior to this study. Table 6 Diagnostic Classification of Subjects in Each Group Diagnosis High Depressed Low Depressed Success Neutral Failure Success Neutral Failure Schizophrenic 1* 1 1 3 3* 4 Psychotic Depression 0 0 0 2 0 Neurotic 3M 2 0 1 1 0 Neurotic Depression 1 2 4 0 0 1 Personality Disorder 2## 3 2 U 2 2# Counseling (Specific Problems) 1# 0 1* 2* 0 1 *History of criminal arrest //History of drug use Information concerning previous psychiatric treatment is presented in Tables 7 and 8. This information was obtained through questioning each patient and was later checked with hospital records. In cases of discrepancies between these two sources, the data in the hospital record were used. As indicated in Table 7, 11 of the 48 subjects had had no previous treatment. Of the remaining 37, the majority had their f i r s t treatment within five years prior to the study. Seven of these 37 had 19 previous treatment on an outpatient basis only. Of the 30 who had at some time been hospitalized, 23 had from one to three hospitalizations. Table 8 also indicates that five patients had from 7 to 10 previous hospitalizations. Of these five, four were from the group of six mentioned earlier who were not new intake patients at the hospital. In addition, two of the six had started psychiatric treatment 22 and 29 years prior to the study whereas only 4 of the 42 new intake patients had started treatment this long ago. These differences raised doubts concern-ing the advisability of including these six patients in the study. However, for reasons which w i l l be discussed later, these six were included. A l l subjects were selected on the basis of the Beck Depression Inventory (DI) (1961, 1967), which may be found in Appendix A. On the basis of each patient's score on this scale, he was categorized as High Depressed or Low Depressed. Although Loeb et al_. (1964) used a range of 0 to 10 for Low Depression and 16 to 63 for High Depression, the present study used scores of 10 and 20 as cut-off points to enable greater differentiation between the two groups. The mean DI score for the Low Depressed patients was 6.37 and for the High Depressed patients 26.37. 20 Table 7 Age at First Psychiatric Treatment and the Number of Years Ago Treatment was First Received for Each Group Group No Previous Number of Years Ago Age at First Treatment Treatment Fir s t Treated 1-5 6-10 11-20 21-30 1-20 21-30 31-40 41-43 High Depressed Success 2 6 0 0 0 3 3 0 0 Neutral 1 4 2 1 0 2 5 0 0 Failure 3 2 1 1 1 1 2 1 1 Mean number of years ago f i r s t treated: 6.55 years. Median number of years ago f i r s t treated: 4.50 years. Mean age at f i r s t treatment: 23.33 years. Median age at f i r s t treatment: 22.00 years. Low Depressed Success 3 2 2 0 1 1 2 2 0 Neutral 1 2 2 2 1 4 1 1 1 Failure 1 4 0 1 2 5 2 0 0 Mean number of years ago f i r s t treated: 10.84 years. Median number of years ago f i r s t treated: 6.00 years. Mean age at f i r s t treatment: 23.84 years. Median age at f i r s t treatment: 20.00 years. Means and Medians based on those who had some previous treatment. 21 Table 8 Frequency of Previous Hospitalization, Outpatient Care, and Shock Treatment for Each Group Group Number of Previous Shock Outpatient Hospitalizations Treatment Care 0 1 2-3 4-6 7-10 Yes No Yes No High Depressed Success 3 2 1 0 0 3 3 4 2 Neutral 3 2 2 0 0 0 7 6 1 Failure 0 1 2 1 1 2 3 3 2 Mean number of previous hospitalizations: 1.55. Median number of previous hospitalizations: 2.00. Low Depressed Success 0 1 2 0 2** 4 1 4 1 Neutral 0 2* 2 2 1* 5 2 7 0 Failure 1 2 3* 0 1* 4 3 4 3 Mean number of previous hospitalizations: 3.42. Median number of previous hospitalizations: 3.00. Based on those who had some previous treatment. *The six subjects who were not new intake patients. Apparatus The Depression Inventory (DI) (Beck, 1961, 1967) consists of 21 multiple-choice items, with each alternative being score 0, 1, 2, or 3. The f i n a l scores range from 0 to 63, the higher scores representing greater depression. The development of this inventory is presented by Beck (1961, 1967), along with research concerning i t s vali d i t y and re-l i a b i l i t y , only part of which w i l l be discussed here. 22 In two patient samples, 200 cases and 606 cases each, Beck (1961, 1967) found the internal consistency of the DI to be quite high, each item correlating at the .01 level or higher with the total score. Split-half r e l i a b i l i t y (odd versus even number items) of 97 cases resulted in a Pearson correlation coefficient of 0.86. In addition, a variation of the test-retest method was employed, whereby changes in the DI score over a 2 to 6 week period were correlated with changes in depth of depression ratings made by a psychiatrist. Results indicated that these changes closely paralleled each other. Beck has also concerned himself with the validity of the DI. The concurrent validity was investigated by means of calculating the Pearson bi s e r i a l correlation between DI scores and c l i n i c a l judgements of depth of depression. In two studies, these were found to be 0.65 and 0.67, both significant at the .01 level (Beck, 1961, 1967). Metcalfe and Goldman (1965), using British patients and psychiatrists' ratings, found the Kendall rank correlation coefficient to be 0.61, significant at the .001 level. Nussbaum, Wittig, Hanlon, and Kurland (1963) found that DI scores had a higher correlation with c l i n i c a l ratings than did Minnesota Multiphasic Personality Inventory protocols. The matching task employed 15 sets of stimulus cards, 2 cards per set. Each card was made of white poster cardboard and measured 9 inches x 12 inches. A l l geometric figures were drawn in black ink. The f i r s t card of each set contained one geometric figure in the center of the card. The second card consisted of this figure and 3 others, numbered 1 through 4. For the second card of each set, the correct alternatives were randomly placed in position 1, 2, 3, or 4 for the 15 cards. The geometric figures 23 used for this task are included in Appendix B. For the line length estimation task, six pieces of white poster cardboard, 9 inches x 12 inches, were used. Of the six lines, two were 8 inches long, one 5 inches, one 4 inches, one 9 inches, and one 2 inches. These lines were drawn in black ink and placed on the six cards at varying angles. To avoid the possible effects of a response bias to over- or under-estimate line lengths, three of the hinted lengths were greater than and three less than the actual lengths. Copies of the six cards are included in Appendix C. Appendix D contains the five-point continuum shown to each patient for the post-task evaluation. This was also made of 9" x 12" white poster cardboard, with the continuum and lettering in black ink. The five points were labeled: "much worse than most other people," "worse than most other people,""the same as most other people," "better than most other people," and "much better than most other people." Procedure A l l patients, other than the six mentioned previously, were seen during their f i r s t v i s i t to the department. During this f i r s t v i s i t , patients were given a standard intake interview, usually by one of two nurses, to obtain information concerning referral source, previous psychiatric treatment, and current d i f f i c u l t i e s . Some of the patients then had a short interview with a staff psychiatrist or resident. Due to the department procedure of allowing intake patients to come anytime between one and three PM, five days a week, without specific appointment, there were often several intake patients waiting or in various stages of the intake procedure simultaneously. For this reason, patients were 24 seen by the experimenter at different points of the intake procedure. This departmental procedure, and the fact that the department i t s e l f gathered DI scores on new patients during part of the time that the present data were collected, created the necessity of a minority of subjects being given the DI by one or the other of two nurses. Although these variables could not be controlled, there is no reason to believe that they are not random for a l l groups of subjects, or that they could have appreciably affected the results, The procedure for administering the DI was always a close variant of the following: Patients were seated at a table, given the DI and a pencil, and told: "This is just a short form we'd like you to f i l l out. Notice that i t consists of groups of statements" (the patient is shown how the statements are grouped). "What you're to do is f i r s t , read through the whole group" (pointing to the f i r s t group) "and then decide which one of the statements describes yourself the best and put a c i r c l e in front of that statement. Mark one statement for every group of statements. Do you have any questions?" The interviewer then l e f t the room while the patient completed the DI. In order to avoid experimenter bias through the interviewer's knowledge of the subject's depression level at the time of the testing, the department receptionist scored a l l DI's and assigned patients to one of the three experimental groups of success, failure, or neutral. The receptionist kept a running account of the number of subjects in each group, enabling her to assign each patient to the group with the 25 currently lowest number of subjects which was also consistent with his depression level and sex. An attempt was thus made to ensure that each group was completed at roughly the same time. However, due to the fact that the majority of patients who came to the department scored at the High Depression level, the High Depressed groups were completed 10 weeks before the Low Depressed groups. In addition, the experimenter was thus aware of the depression level for the last 11 of the 24 subjects of the Low Depressed groups. After being assigned to one of the three experimental groups, the experimenter and the subject entered a small interviewing room and sat opposite each other at a table. Patients were then instructed as follows: "I have a few things I'd like you to do. The f i r s t thing I want you to do is to estimate some intervals of time. I want you to estimate how much time elapses between the point at which I say start and when I later say stop. Do this without counting out loud or in your head. We w i l l do this six times. Do you have any questions? Remember, a l l I want you to do is t e l l me how much time elapses between the words start and stop." The experimenter then said, "Ready, start," and after the allotted time, "Stop," and asked for the subject's estimate. For a l l patients, the six t r i a l s were 20 seconds, 70 seconds, 50 seconds, 10 seconds, 90 seconds, and 30 seconds, in that order. The time intervals were measured with a stop watch and each estimate was recorded. For the no-feedback group, the only comments made between the t r i a l s were comments such as " A l l right, let's do the next one, ready?" Patients in the failure feedback condition were told the following state-ments after each t r i a l : 26 1. "That wasn't very good, see i f you can do better on the next one." 2. "Hm" (repeat estimate and look at watch), "well, let's try another one." 3. "That s t i l l wasn't very close, but let's see how you can do on this one." 4. "You should be closer than that. Maybe I should read the instructions again for you. Remember, a l l I want you to do is t e l l me how much time elapses between the words start and stop. Okay?" 5. "That's way off, let's try the last one." 6. (repeat his estimate) "I thought you would have done better than that. It was the last one though so maybe you can do better on something else." Success feedback subjects were told: 1. "Very good. That was quite close." 2. "Excellent. That was a very accurate estimate." 3. "Good. You can't get much closer than that." 4. "You must have a clock in your head or something. I don't know how you are getting them so close." 5. "I can't believe how accurate you are." 6. "Very good. That was the last one. You did extremely well on these." The instructions for the matching task and performance predictions were as follows: "Now I have something else for you to do. I'm going to show you a card on which there is a figure drawn. I want you to look at i t for a few seconds and then I w i l l take i t away. When I take i t away, I ' l l show you another card which has four figures on i t , numbered one through four, one of which w i l l be exactly lik e the one on the f i r s t card but the other three w i l l not be exactly like i t . I want you to look at the f i r s t figure carefully so that when I take i t away and show you the second card you can t e l l me the number of the figure which is exactly like the one on the f i r s t card. Okay? Do you understand what I want you to do? Any questions? Now, we w i l l dc this same thing for 15 different sets of cards, so your f i n a l score of right answers could be anything from 0 to 15. Before we 27 begin, I want you to t e l l me how many of the 15 you think you w i l l get right. If there were a stranger here with us, someone who has never seen you before, how many points would he think you would be able to get? What i f I asked your mother? And your father? I ' l l begin showing you the cards now. This is the f i r s t one." The 15 sets of stimulus cards are included in Appendix B, in the order in which they were presented to each patient. Subjects were allowed to view the f i r s t card for five seconds. They were then shown the second card and asked to select the figure which was identical to the stimulus figure. Their predictions and answers on the matching task were recorded. Since the last task, that of estimating the lengths of six lines, was used as a measure of conformity behavior, the instructions to patients involved pressure to change their answer to an erroneous one. The instructions were as follows: "The last thing I want you to do is to estimate the length of six lines. I w i l l show you the lines one at a time, each drawn on a different piece of paper. A l l you have to do is to look at the line briefly and t e l l me how long you think i t i s . To help you make a good decision I w i l l give you a hint about the length of each line before I show i t to you. I can't t e l l you the exact length of the line, but I w i l l t e l l you the estimate most often given by other people I've asked to do this. Do you understand what I want you to do?" Each card was presented to a l l subjects in the order in which they are presented in Appendix C. Before each card was shown to the patient, he was told the hinted length, also included in this Appendix. When 28 giving the hints, the interviewer said, "The estimate most often given for this one i s inches," or "Most people guess this one to be about inches." A l l line estimates were recorded. The instructions for the post-task evaluation were as follows: "And one other thing I'd like you to do. It concerns how well you think you did, compared to other people, on the matching task [the subject is shown the continuum scale]. You see this line with five points. I want you to t e l l me which of these five points describes the way you think you did on the figure matching. Do you think you did much worse than most other people, worse than most other people, the same as most other people, better than most other people, or much better than most other people?" The continuum used is presented in Appendix D. Each subject's position score was recorded. Since this measure asked patients to compare their performance with that of other people, i t was not included prior to the line estimation task due to possible confounding with conformity behavior. Just prior to leaving, information was obtained concerning each patient's age, educational level, occupation, parents' educational level, and previous psychiatric treatment. A l l subjects were thanked and assured that they performed well. The patients were given no explanation for their participation in the study; i t was treated as part of the hospital procedure. CHAPTER THREE RESULTS To determine whether the six subjects who were not new intake patients should be included in the data analyses, a comparison of the means on a l l the dependent variables for the total 48 subjects and the 42 who were new intake patients was carried out. These data are presented in Appendix E. The comparison was carried out between the 42 and 48' subjects, rather than by comparing the means of the 6 with the other 42, since not a l l groups have members from the group of 6 subjects, and thus some cells would have no entries, preventing a complete comparison of any differences between group means. This comparison indicated that the inclusion of these 6 patients did not appreciably affect the means nor the direction of differences between the means on any of the dependent variables. Although time estimation was not a dependent variable, the means for this task are also included in Appendix E. While some of these means did change as a result of including these 6 patients, i t was f e l t that this did not warrant their elimination, as this was the feedback task, not a variable. Therefore, the analyses which follow were carried out using a l l 48 subjects. To investigate the hypotheses concerning the post-task evaluation, a three-factor analysis of variance, sex x depression level x experimental condition, was carried out for the position scores, 1 to 5. This analysis yielded no significant differences. Table 9 contains the mean position score for each group. As shown in the table, a l l means for the post-task evaluation were quite similar, with the majority of patients evaluating their performance at position 3, "the same as most other people." 30 Table 9 Means of the Post-Task Evaluation Success Neutral Failure High Depressed Males 3.50 3.25 3.50 Females 3.00 3.25 3.75 Low Depressed Males 3.25 2.75 3.75 Females 3.25 3.00 3.25 Before investigating the hypotheses concerning conformity behavior, i t was f i r s t necessary to show that errors made on this task were not random, but that the task and instructions were effective in producing more errors in the hinted direction than in the nonhinted direction. To do this, a _t test was carried out between the total number of inches of error in the hinted direction and the total number of inches of error in the nonhinted direction over a l l six items for a l l subjects. These means, and their breakdown by depression level, are shown in Table 10. As the table shows, the mean number of inches of error in the hinted direction for a l l items and a l l subjects was significantly larger than the mean number of inches of error in the nonhinted direction. When collapsing across a l l six items for each patient, the total amount of inches of error in the hinted direction was greater than the total inches of error in the nonhinted direction for 43 of the 48 subjects. The 5 patients 31 whose nonhinted error was greater were a l l High Depressed subjects. Table 10 Mean Inches of Error in the Hinted and Nonhinted Direction for a l l Conformity Items Group Hinted Direction Nonhinted Direction 48 Subjects* .8775 .2786 High Depressed .7065 .2725 Low Depressed 1.0486 .2847 *_t value significant beyond .01 level, using 2-tailed test. To investigate the hypotheses concerning conformity behavior, a three-factor analysis of variance was carried out. The score used in this analysis was the total amount of inches of error in the hinted direction for a l l items, with errors in the nonhinted direction and accurate responses scored as zero. Table 11 contains a summary of this analysis and Table 12 presents the mean inches of error in the hinted direction for each group. The only significant effect found in the analysis was for depression level, with the High Depressed patients having significantly fewer inches of error in the hinted direction than the Low Depressed ones, with means of 4.23 and 6.31, respectively. Since this analysis was based on the amount of error in the hinted direction only, i t was considered necessary to further analyze the data in terms of the frequency of responses at each depression level which were Table 11 Analysis of Variance of the Magnitude of the Conformity Response Source df MS F Sex (S) 1 6.5638 . o. 8192 Experimental Condition (E) 2 15.6810 1. 9572 Depression Level (D) 1 51.5638 6. 4360* SxE 2 7.0013 0. 8738 SxD 1 19.0638 2. 3794 ExD 2 0.8294 0. 1035 SxExD 2 0.0091 0. 0011 *£<.05 Table 1 2 Mean Inches of Error in the Hinted Direction on A l l Conformity Items Success Neutral Failure High Depressed Males 3.62 4.56 5.31 Females 3.75 2.56 5.62 Males Females 4.12 6.75 Low Depressed 5.93 6.37 5.87 8.81 33 accurate, in error in the hinted direction, or in error in the nonhinted direction, regardless of the magnitude of error. Table 13 presents the frequency of each type of response for each group. As the table shows, Table 13 Frequency of Each Type of Response for Both Depression Levels Group Accurate Hinted Nonhinted Direction Direction High Depressed 42 71 31 Low Depressed 30 88 26 High Depressed subjects gave more correct responses, slightly more in the nonhinted direction, and fewer in the hinted direction. Chi square analysis of these frequencies indicated that these differences were not significant. Thus, Low Depressed patients did not give significantly more conformity responses, but their conformity responses were more in error than were those of the High Depressed subjects. It is therefore possible that the significance reported earlier i s primarily due to the fact that the responses of Low Depressed patients were more in error in both the hinted and nonhinted direction than were the responses of High Depressed patients. However, as Table 10 indicated, the means of the number of inches of error in the nonhinted direction for High and Low Depressed subjects were .2725 and .2847, respectively. Therefore, while Low Depressed patients gave a greater magnitude of error in the hinted 34 direction, the magnitude of error of their responses in the nonhinted direction was no greater than that of High Depressed subjects. The inter-item r e l i a b i l i t y of the six conformity items was investi-gated through chi square analysis of the frequencies of subjects who gave answers in error in the hinted direction and answers i n error in the nonhinted direction for each item. These frequencies are presented in Table 14, along with the percentage of nonaccurate subjects who gave each type of response for each item. This chi square value was s i g n i f i -cant beyond the .01 leve l , indicating that some items were more effective in producing conformity responses than were others. As shown in Table 14, Table 14 The Frequencies and Percentages of Both Types of Nonaccurate Responses for Each Item Item 1 2 3 4 5 6 Responses in Hinted Direction Frequency of Subjects 29 28 19 17 45 21 Percentage of Subjects 82.9 80.0 73.1 45.9 97.8 56.8 Responses in Nonhinted Direction Frequency of Subjects 6 7 7 20 1 16 Percentage of Subjects 17.1 20.0 26.9 54.1 2.2 43.2 items 4 and 6, both giving hinted lengths less than the real lengths, were the least effective i n producing errors in the hinted direction, as opposed to errors in the nonhinted direction. Although the data do not 35 enable the investigator to determine the reason for the ineffectiveness of these two items, this could not have accounted for the results reported earlier as the responses of neither group were particularly affected by them. Of the subjects who gave non-conforming responses on both these items, exactly one-half were High Depressed and one-half were Low Depressed. The four performance predictions were analyzed in a four-factor repeated measure analysis of variance. The measure used was the prediction socre, 0 to 15. This analysis i s summarized in Table 15 and the means for each prediction and each group are presented in Table 16. As Table 15 indicates, the only significant main effect found was the performance prediction. Simple effects analysis of this factor, for which means are presented in Table 17, indicated that self and mother predictions were significantly different, with means of 9.70 and 10.70 respectively. Stranger and mother predictions also differed significantly, with means of 9.29 and 10.70. Thus, maternal performance prediction was significantly greater than either self or stranger predictions, but performance pre-dictions were unrelated to the level of depression. A significant three-factor interaction was found for Sex x Experimental group x Prediction. Although simple effects analysis of this interaction indicated that these differences were, in some cases, opposite that expected, i t does suggest that the experimental feedback had at least a minimal effect for some subjects. The means for this simple effect analysis, presented in Table 18, indicated that male and female predictions differed significantly only for the maternal prediction, in the success condition, for which males gave a significantly 36 Table 15 Analysis of Variance of the Four Performance Predictions Source df MS F Sex (S) 1 18.75 0.7596 Experimental Condition (E) 2 36.78 1.4902 Depression Level (D) 1 27.00 1.0938 S x E 2 32.95 1.3349 S x D 1 0.02 0.0008 E x D 2 53.07 2.1503 S x E x D 2 28.97 1.1737 Predictions (P) 3 17.07 3.0898* S x P 3 6.25 1.1308 E x P 6 4.86 0.8798 D x P 3 3.00 0.5428 S x D x P 3 7.29 1.3206 S x E x P 6 18.39 3.3276# E x D x P 6 3.93 0.7115 S x E x D x P 6 1.68 0.3056 *£<.05 #£<.01 37 Table 16 Means for the Four Performance Predictions for Each Group Group Self Stranger Mother Father Males Success High Depressed 9.00 8.50 11.25 8.75 Low Depressed 10.75 11.50 12.50 12.00 Neutral High Depressed 12.25 9.50 11.00 10.75 Low Depressed 9.50 10.00 8.00 6.00 Failure High Depressed 9.00 7.75 10.50 10.00 Low Depressed 12.25 10.75 12.50 11.75 Females Success High Depressed 7.25 8.50 5.50 7.00 Low Depressed • 11.00 8.75 8.75 9.50 Neutral High Depressed 8.00 7.75 11.75 10.50 Low Depressed 8.25 8.00 11.75 10.25 Failure High Depressed 8.75 10.75 13.25 12.00 Low Depressed 10.50 9.75 11.75 11.50 Mean High Depressed 9.04 8.79 10.54 9.83 Mean Low Depressed 10.37 9.79 10.87 10.16 38 Table 17 Means for the Four Performance Predictions for A l l Subjects Prediction Mean Self (S) 9.70 Stranger (T) 9.29 Mother (M) 10.70 Father (F) 10.00 £<.01 for S versus M, and for T versus M. Table 18 Mean Performance Predictions for the Significant Three-Factor Interaction Group Prediction Self Stranger Mother Father Males Success 9.87 10.00 11.87 10.37 Neutral 10.87 9.75 9.50 8.37 Failure 10.62 9.25 11.50 10.87 Females Success 9.12 8.62 7.12 8.25 Neutral 8.12 7.87 11.75 10.37 Failure 9.62 10.25 12.50 11.75 39 higher prediction than did females. In no case were male performance predictions significantly altered due to experimental group; however, for females, both mother and father predictions were altered due to experi-mental condition. For females, the maternal prediction was significantly lower in the success condition than in the neutral or failure conditions. The paternal prediction was significantly lower in the success condition than the failure condition. Thus, the success condition was more effective than the failure condition in changing the maternal prediction for females. In addition, there was greater va r i a b i l i t y between the four predictions within the neutral and failure conditions for females than for males in any condition. The success condition tended to eliminate some of this v a r i a b i l i t y among the four predictions for females, by bringing the four predictions closer together. However, even in the success condition, female self-prediction was significantly higher than female maternal prediction. In the neutral and failure conditions, maternal predictions for females were significantly higher than self-predictions. Thus, for females, maternal and self-predictions differed in every condition, while paternal and self-predictions differed only in the neutral and failure conditions, with the self-prediction being lower than the paternal prediction in each case. In no case, for either males or females, did maternal and paternal predictions differ significantly. For males, the self-prediction was significantly lower thatf the maternal prediction in the success condition (the opposite of what was found for females), and the self-prediction was significantly higher than the paternal prediction in the neutral condition (again the opposite of what was found for females). 40 In some cases the stranger prediction differed significantly from the maternal and paternal predictions. The stranger prediction was significantly different from the maternal prediction for males in the failure condition, and for females in the neutral and failure conditions, and from the paternal prediction for females in the failure condition.; In each case the stranger prediction was lower than the maternal or the paternal predictions. Although not a dependent variable, the total number of errors made on the matching task was also investigated by a three-factor analysis of variance. Table 19 contains the mean number of errors made for each group. Although the differences were not significant, the combined means indicated that there were slightly fewer errors made by subjects in the neutral condition, and fewer made by High Depressed patients than by Low Depressed patients. Table 19 Mean Errors Made on the Matching Task Success Neutral Failure High Depressed Males 3.75 5.25 3.50 Females 3.25 2.50 5.50 Males Females 5.25 5.50 Low Depressed 4.25 3.25 3.75 5.75 CHAPTER FOUR DISCUSSION There were two kinds of hypotheses made in the present study: (a) hypotheses concerning differences between High and Low Depressed subjects on the variables investigated in the neutral condition, without any experimental feedback, and (b) hypotheses concerning the differential effect of performance feedback on High and Low Depressed subjects on the variables investigated. None of the hypotheses of either type were supported by the present data. The only measure significantly related to depression level was the magnitude of the conformity response, for which the results were opposite that expected. There was no difference between depression levels on the number of conformity responses made, nor was there a differential effect of experimental feedback on level of depression for the variables investigated. The general lack of hypothesized differences between depression levels may have been due to ineffectiveness of the Beck DI in differentiating between High and Low Depressed subjects; or, limitations of Beck's theory in predicting the behavior of depressed patients in the present experimental situation. The lack of differences between subjects in the three experimental conditions was more li k e l y due to the nature of the experimental situation. Possible Limitations of Beck DI The lack of expected differences between depression levels may have been due to the use of the Beck DI as the only criterion in determining the patients' level of depression. While the investigator believed that i t was best to avoid the use of hospital diagnostic classification in the selection of subjects, i t also became apparent 42 that the use of the DI alone was unsatisfactory in differentiating between depressed and nondepressed patients. While the r e l i a b i l i t y and validity of the DI have been investigated with often promising results, i t i s possible that the DI has too large a frequency of false positives and false negatives to be used as the only selection criterion especially when investigating etiological variables of depression. Some patients may have responded at the high depression level on the DI primarily as a plea for help, while low scores, in some cases, may have been due to defensiveness in admitting more pathological self-descriptions. Further, items on the DI refer primarily to current characteristics of depression, while some of the hypotheses investigated in the present study related primarily to Beck's etiological picture of depression. If depression is for many patients cy c l i c a l in nature, their scores on the DI would also be cyc l i c a l over time, although occurrences in their early development would remain constant. Thus, a patient whose early developmental years were consistent with Beck's etiological picture of depressives could score as High Depressed on the DI at one point, and as Low Depressed at another, suggesting that this scale may be of limited value when studying etiological variables of depression. Possible Limitations of the Experimental Manipulation The lack of differences between a l l subjects in the three experi-mental conditions may have been due to the use of time estimation as the experimental feedback task. Unlike the task of word completion, used by Loeb et a l . (1964), accuracy in time estimation was possibly not viewed as an important a b i l i t y nor as an indicator of intelligence. Thus, feedback on this task may have been ineffective in providing subjects 43 with a meaningful experience of success or failure. The patients' perception of the experimenter and the testing situation may have also decreased the effectiveness of the experimental feedback. Patients were seen during their f i r s t v i s i t to the department and i t is possible that the attitude of seeking help at this time was so strong that a l l those who came into contact with the patient were perceived primarily as helpers, and the situation as a means of gaining help. This perception of the experimenter and the situation may have limited the subjects' a b i l i t y to experience the situation and performance feedback as a success or failure. It is also possible that patients at this time were seeking an explanation of their d i f f i c u l t i e s , as one does when seeing a physician for physical problems. If this was the case, the performance feedback may have been perceived more as a physician's comment concerning one's health, than as an evaluation of one's a b i l i t y . Thus, rather than experiencing the situation as one of a personal success or failure, patients may have experienced that which they wanted and expected: possible help and an explantion of their d i f f i c u l t i e s . Possible Limitations of Beck's Theory of Depression If we assume that the DI did provide a valid and meaningful separation of patients in the present study, the lack of findings may necessitate a reconsideration of Beck's theory of depression. While the following speculations are consistent with Beck's theory, they indicate possible limitations of i t , as well as some changes in emphasis. Beck's description of depression emphasizes the depressive's defeatism, self-devaluation, and pessimistic attitude. Other researchers have emphasized the depressive's dependency on the approval and security. 44 of others as motivating forces for his behavior, and his inability to use and be motivated by an internal standard (Becker, 1960; Katkin, Sasmor, & Tan, 1966). Thus, while Beck's theory implies that depressed patients, as compared to nondepressed patients, would give more negative performance evaluations and predictions in any situation, the social dependency hypothesis suggests that this may only be the case in a situation where there can be some question as to one's social status and approval, or on a task that is directly related to one's social status. The lack of findings in the present nonsocial situation suggests that some changes in Beck s theory may be needed, such that i t allows for the depressive's social dependency and i n a b i l i t y to work towards the satisfaction of internal standards. The following discussion attempts to describe the development of this social dependency, in terms of Beck's etiological theory. From what has already been said concerning depressives, i t i s possible to hypothesize that rather than having internalized any specific standards or idea of an "other," depressives have only internalized the idea that they must win the approval of others, whereas nondepressives are able in their development to select from the mass of others those whose approval they want to seek and gain, and thereby internalize an image of this specific significant "other," the family background studies suggest that depressives, through parental training, may have developed a need to seek and win the approval of a l l others. To them, a l l others may be significant "others."' The heterogeneity of this massive significant "other" would prohibit the development and internalizations of a clearly defined image of significant others. Further, while nondepress-45 ives would be able to differentiate the standards of evaluation of their significant others and thus develop and internalize an evaluative standard of their own, depressives would receive many heterogeneous standards from their more inclusive significant others, and thus be unable to clearly define and internalize any specific evaluative standard. Thus, nondepressives may possess an internalized image of their significant "other" and an internalized evaluative standard, which would provide motivation for their achievement behavior and also allow them to evaluate and label their own performance in the absence of significant others or cues from others. Depressives, on the other hand, may not possess this type of internalized "other" or internalized standard, but may instead seek to satisfy a l l others, on a l l standards. The d i f f i c u l t y of such a task, and the depressive's identification with parental devaluation of himself, may result in the triad of depressive characteristics described by Beck. Further, lacking clear internal standards, the motivation for the depressive's achievement behavior would be primarily social status and the security and approval of others. This external motivation and dependency on others' standards suggest that the depressive would have d i f f i c u l t y evaluating his own performance in any specific situation, and in fact, that specific self-evaluation may be relevant to the depressive only in a social situation. However, i f the depressive, as Beck suggests, has internalized a negative parental evaluation of himself and is especially sensitive to and afraid of experiencing failure, he may often assume he has failed at winning the approval of others in specific social situations. This assumption of failure, however, may not be a self-evaluation on the part of the depressive, but an assumption of the 46 specific external other's evaluation of the depressive. Thus, whereas Beck suggests that the depressive has identified with and internalized a parental devaluation of himself, the present suggestion is that the depressive has internalized (a) parental emphasis on winning social status and approval, and (b) parental devaluation of his a b i l i t y to do this. Beck's theory implies that depressives would devaluate their performance in any situation. The present hypothesis, however, suggests that this may only occur i f the evaluation is relevant to the depressive's a b i l i t y to win social approval, and further, that depressives may be un-concerned with their actual a b i l i t y or performance, as anything other than a means of gaining social reinforcement, the approval of others. The lack of findings in the present nonsocial situation clearly does not support Beck's theory, nor can i t be used as direct support of the social dependency hypothesis. However, a consideration of the present results, Loeb et al.'s (1964) results concerning performance predictions in a competitional setting, and Katkin et al.'s (1966) results concerning 1 conformity behavior on a social opinion task, suggests that the social dependency hypothesis may supply a useful addition to and modification of Beck's theory. CHAPTER FIVE SUMMARY AND FUTURE DIRECTIONS The present paper reported the results of a study investigating the differential effect of success and failure performance feedback on psychiatric patients varying in level of depression. Based on Beck's descriptive and etiological formulation of depression and the results of earlier studies, i t was expected that High Depressed subjects, compared to Low Depressed subjects, would: (a) Give a lower prediction of their future performance on a specific task. (b) Give a lower parental and stranger prediction of their future performance. (c) Give a lower post-task evaluation of their performance. (d) Show more conformity behavior. Finally, i t was hypothesized that: (e) Success would raise the predictions and post-task evaluation, and lower the amount of conformity behavior for a l l subjects, failure having the opposite effect, but that High Depressed subjects would be more affected by the performance feedback than Low Depressed subjects. These differences were investigated using a total of 48 outpatient psychiatric patients, 24 who scored at the High Depression level and 24 at the Low Depression level on the Beck Depression Inventory (DI) (Beck, 1961, 1967). The experimental situation involved an experimentally induced success, failure, or neutral experience through a task of time estimation, during which patients received feedback regarding their performance consistent 48 with their experimental group. Following this, a l l patients were asked to give the performance predictions lis t e d above for a task of matching geometric figures. The conformity situation involved estimating the lengths of six lines, with subjects being given a hinted length for each. Each patient was then asked to evaluate his performance on the matching task, using a 5-point scale. None of the hypotheses concerning differences between depression levels, nor the differential effect of performance feedback, were supported by the data. It was suggested that the lack of expected differences between the three experimental groups may have been due to the type of experimental feedback task used, or the manner in which patients perceived the inter-viewer and the experimental situation. Ineffectiveness of the Beck DI in differentiating between depression levels may account for the lack of differences between High and Low Depressed subjects on the variables investigated. Certainly future research requires better techniques for classifying patients as depressed or nondepressed. The meaningful use of questionnaire scale scores, such as DI scores, i s confounded by those variables, mentioned earlier, which f i l t e r and dictate the material a patient verbally reports. While factor analytic studies may prove useful in exploring other dimensions of depression, and thus providing more effective tools for this purpose, in the meantime i t may be appropriate to employ more than a single test score as the criterion of depression level. It was also suggested that, i f the DI was effective in differenti-ating between depression levels, the present lack of results may 49 necessitate a modification of Beck's theory of depression. While Beck has emphasized the depressive's defeatism, self-devaluation, and rejection, other researchers have emphasized the depressive's dependency on social approval and security as motivating much of his behavior. The lack of findings in the present study may be due to the failure of a non-social situation to provide the motivation for depressed patients' conformity and achievement behavior. This social emphasis implies that the cognitive triad Beck has suggested to be characteristic of depressed patients may be limited to what the depressive perceives to be socially relevant aspects of his l i f e . While the depressive may devaluate himself in general, this may, to the depressive, actually be a social devaluation, since the ambiguity of global evaluation allows him to perceive i t in terms of his a b i l i t y to win social approval. Specific self-evaluation may only be of concern to the depressive in what he perceives to be a socially relevant situation. When the depressive does devaluate his performance in a specific social situation, he may not be devaluating his a b i l i t y on the task performed, but his a b i l i t y to perform well enough to win social approval. It was also suggested that this social dependency and devaluation may be due to an internalized parental emphasis on social approval, and devaluation of the depressive's abi l i t y to achieve this approval. Thus, as Beck suggests, the depressive may have learned from significant others that he i s inferior, but i t is further suggested that he has learned only that he is socially inferior, and in fact, that social a b i l i t y i s a l l that i s important. Although this interpretation of the depressive's functioning is 50 consistent with Beck's theory and the results of earlier studies, the present emphasis upon social dependency, rather than self-devaluation, limits the situations and areas in which commonly accepted depressive characteristics may be found. While earlier studies found significant differences between High and Low Depressed subjects on variables similar to those presently investigated, in each case the situation was social, the task socially relevant, or the requested self-evaluation nontask specific, allowing the depressive to perceive the evaluation as social. It is therefore necessary to investigate the depressive's performance on tasks varying in degree of social relevancy, and in social and nonsocial situations. This would involve a comparison of depressives' performance in situations similar to that of the present design, in observational situations, and in competitional situations, such as used by Loeb et a l . (1964). Future investigations might also compare their performance on perceptual judgement tasks and on Katkin et_ al.'s (1966) social opinion task. A more direct investigation of these variables, however, would involve studying the family structure and development of children who later become depressives. If the present hypotheses prove supportable in future research, i t is possible that training the depressive to be able to evaluate himself and his performance in terms other than his a b i l i t y to win social status and approval, where he presumably i s not necessarily self-deprecatory, and emphasizing the importance of nonsocial evaluation, may help counteract and decrease the effectiveness of his internalized social devaluation. 51 FOOTNOTES 1. The author wishes to express thanks and gratitude to Dr. P. Termansen and Dr. H. Klonoff, and a l l the staff at Vancouver General Adult Out-patient Department of Psychiatry for their cooperation and help in collecting the data. The author is especially indebted to Miss Nancy Trites, department receptionist, for her kind assistance throughout the data collection. The author also wishes to express thanks to Dr. D. Papageorgis, Dr. S. Butt, and Dr. A. Marlatt, without whose help and guidance this project would not have been possible. REFERENCES Adorno, T. W., Frenkel-Brunswik, E., Levinson, D. J., & Sanford, R. N. The authoritarian personality. New York: Harper, 1950. Arthur, A. Z. Response bias in the semantic dif f e r e n t i a l . British Journal of Social and C l i n i c a l Psychology, 1966, 5_, 103-107. Beck, A. T. Depression: C l i n i c a l , experimental, and theoretical aspects. New York: Harper, 1967. Beck, A. T. & Stein, D. The self-concept in depression. Unpublished study, 1960. Cited by Beck, 1967, p. 164. Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. An inventory for measuring depression. Archives of General Psychiatry, 1961, 4^, 561-571. Becker, J. Achievement-related characteristics of manic-depressives. Journal of Abnormal and Social Psychology, 1960, 60, 334-339. Becker, J., Spielberger, C. D., & Parker, J. B. Value achievement and authoritarian attitudes in psychiatric patients. Journal of C l i n i c a l Psychology, 1963, 19, 57-61. Brunning, J. & Kintz, B. Computational handbook of s t a t i s t i c s . 111.: Scott, Foresman, & Co., 1968. Cohen, M. B., Baker, G., Cohen, R. A., Fromm-Reichmann, F., & Weigert, E. V. An intensive study of twelve cases of manic-depressive psychosis. Psychiatry, 1954, J7, 103-137. DeCharms, R., Morrison, W. H., Reitman, W., & McClelland, D. C. Behavioral correlates of directly and indirectly measured achievement motivation. In D. C. McClelland (Ed.), Studies in motivation. New York: Appleton-Century-Crofts, 1955. Dominion Bureau of Statistics. Canadian St a t i s t i c a l Review, 1968, 43, iv. Gibson, R. W. Comparison of the family background and early l i f e experience of the manic-depressive and schizophrenic patient. Final Report on Office of Naval Research Contract (Nonr-751(00)). Washington, D. C , 1957. Cited by Beck, 1967, p. 162. Hildreth, H. M. A battery of feeling and attitude scales for c l i n i c a l use. Journal of C l i n i c a l Psychology, 1946, 2, 214-221. Katkin, E. S., Sasmor, D. B., & Tan, R. Conformity and achievement-related characteristics of depressed patients. Journal of Abnormal and Social Psychology, 1966, 71_, 407-412. 53 Laxer, R. M. Self-concept changes of depressive patients in general hospital treatment. Journal of Consulting Psychology, 1964, 28, 214-219. Levinson, D. J. & Huffman, P. E. Traditional family ideology and i t s relation to personality. Journal of Personality, 1955, 23_, 251-273. Loeb, A., Beck, A. T., Diggory, J. C , & T u t h i l l , R. The effects of success and failure on mood, motivation, and performance as a function of predetermined level of depression. Unpublished study, 1966. Cited by Beck, 1967, p. 161. Loeb, A., Feshback, S., Beck, A. T., & Wolf, A. Some effects of reward upon the social perception and motivation of psychiatric patients varying in depression. Journal of Abnormal and Social Psychology, 1964, 68, 609-616. McClelland, D. C , Atkinson, J. W., Clark, R. S., & Lowell, E. L. The achievement motive. New York: Appleton-Century-Crofts, 1953. Metcalfe, M. & Goldman, E. Validation of an inventory for measuring depression. British Journal of Psychiatry, 1965, 111, 240-242. Myers, J. Fundamentals of experimental design. Boston: Allyn & Bacon, Inc., 1966. Neuringer, C. Effect of intellectual level and neuropsychiatric status on the diversity and intensity of semantic differential ratings. Journal of Consulting Psychology, 1963, 27, 280. Nussbaum, K., Wittig, B. A., Hanlon, T. E., & Kurland, A. A. Intravenous nialamide in the treatment of depressed female patients. Comprehensive Psychiatry, 1963, 4_, 105-116. Osgood, E. E., Suci, G. J., & Tannenbaum, P. H. The measurement of meaning. Urbana, 111.: University of 111. Press, 1957. Roe, A. The psychology of occupations. New York: John Wiley & Sons, Inc., 1956. Spielberger, C. D., Parker, J. B., & Becker, J. Conformity and achievement in remitted manic-depressive patients. Journal of Nervous and Mental Disease, 1963, 137, 162-172. Wilson, D. C. Families of manic-depressives. Diseases of the Nervous System, 1951, 12, 362-369. Winer, B. J. Stat i s t i c a l principles in experimental design. New York: McGraw-Hill, 1962. Wittenborn, J. R. The dimensions of psychosis. Journal of Nervous and Mental Disease, 1962, 134, 117-128. 54 M., Gardiner, D., & Lowy, D. Extreme response tendency as a function of emotional adjustment. Journal of Abnormal and Social Psychology, 1964, 69, 654-657. APPENDIX A The Deck Depression Inventory with the Score Value for Each Alternative 56 A U I do not feel sad. 1 I feel blue or sad. 2a I am blue or sad a l l the time and I can't snap out of i t . 2b I am so sad or unhappy that i t is very painful. 3 I am so sad or unhappy that I can't stand i t . B 0 I am not particularly pessimistic or discouraged about the future. la I feel discouraged about the future. 2a I feel I have nothing to look forward to. 2b I feel that I won't ever get over my troubles. 3 I feel that the future i s hopeless and that things cannot improve. C 0 I do not feel like a failure. 1 I feel I have failed more than the average person. 2a I feel I have accomplished very l i t t l e that i s worthwhile or that means anything. 2b As I look back on my l i f e a l l I can see is a lot of failures. 3 I feel I am a complete failure as a person (parent, husband, wife). D 0 I am not particularly dissatisfied. la I feel bored most of the time. lb I don't enjoy things the way I used to. 2 I don't get satisfaction out of anything any more. 3 I am dissatisfied with everything. E 0 I don't feel particularly guilty. 1 I feel bad or unworthy a good part of the time. 2a I feel quite guilty. 2b I feel bad or unworthy practically a l l the time now. 3 I feel as though I am very bad or worthless. F 0 I don't feel I am being punished. 1 I have a feeling that something bad may happen to me. 2 I feel I am being punished or w i l l be punished. 3a I feel I deserve to be punished. 3b I want to be punished. G 0 I don't feel disappointed in myself. la I am disappointed in myself. lb I don't like myself. 2 I am disgusted with myself. 3 I hate myself. 57 H 0 I don't feel I am any worse than anybody else. 1 I am very c r i t i c a l of myself for my weaknesses or mistakes. 2a I blame myself for everything that goes wrong. 2b I feel I have many bad faults. I 0 I don't have any thoughts of harming myself. 1 I have thoughts of harming myself but I would not carry them out. 2a I feel I would be better off dead. 2b I have definite plans about committing suicide. 2c I feel my family would be better off i f I were dead. 3 I would k i l l myself i f I could. J 0 I don't cry any more than usual. 1 I cry more now than I used to. 2 I cry a l l the time now. I can't stop i t . 3 I used to be able to cry but now I can't cry at a l l even though I want to. K 0 I am no more irr i t a t e d now than I ever am. 1 I get annoyed or irrita t e d more easily than I used to. 2 I feel i r r i t a t e d a l l the time. 3 I don't get irrita t e d at a l l at the things that used to i r r i t a t e me. L 0 I have not lost interest in other people. 1 I am less interested in other people now than I used to be. 2 I have lost most of my interest in other people and have l i t t l e feeling for them. 3 I have lost a l l my interest i n other people and don't care about them at a l l . M 0 I make decisions about as well as ever. 1 I am less sure of myself now and try to put off making decisions i 2 I can't make decisions any more without help. 3 I can't make any decisions at a l l any more. N 0 I don't feel I look any worse than I used to. 1 I am worried that I am looking old or unattractive. 2 I feel that there are permanent changes in my appearance and they make me look unattractive. 3 I fee l that I am ugly or repulsive looking. 58 0 0 I can work about as well as before. la It takes extra effort to get started at doing something, lb I don't work as well as I used to. 2 I have to push myself very hard to do anything. 3 I can't do any work at a l l . P 0 I can sleep as well as usual. 1 I wake up more tired in the morning than I used to. 2 I wake up 1-2 hours earlier than usual and find i t hard to get back to sleep. 3 I wake up early every day and can't get more than 5 hours sleep. Q 0 I don't get any more tired than usual. 1 I get tired more easily than I used to. 2 I get tired from doing anything. 3 I get too tired to do anything. R 0 My appetite i s no worse than usual. 1 My appetite is not as good as i t used to be. 2 My appetite is much worse now. 3 I have no appetite at a l l any more. S 0 I haven't lost much weight, i f any, lately. 1 I have lost more than 5 pounds. 2 I have lost more than 10 pounds. 3 I have lost more than 15 pounds. T 0 I am no more concerned about my health than usual. 1 I am concerned about aches and pains or_ upset stomach or constipation or other unpleasant feelings i n my body. 2 I am so concerned with how I feel or what I feel that i t ' s hard to think of much else. 3 I am completely absorbed in what I fee l . U 0 I have not noticed any recent change in my interest in sex. 1 I am less interested in sex than I used to be. 2 I am much less interested in sex now. 3 I have lost interest in sex completely. APPENDIX B Stimulus Cards Used in the Matching Task 60 The following 15 sets of stimulus cards are presented i n the same order in which they were presented to each subject. The correct responses for each set follow: 1. 1 9. 4 2. 3 10. 2 3. 4 11. 1 4. 4 12.. 4 5. 2 13. i 6. 3 14. 3 7. 1 15. 2 8. 3 61 i-63 6k 65 66 > o > o c > • > o • > • o 67 68 73 i i i 77 79 6 - 80 « i 4 81 i i i • j I 8 3 8 5 86 j i 87 : 89 I i i i i | i > I i 90 • o > o APPENDIX C The Six Line Stimulus Cards Used for the Conformity Task 92 The following six stimulus cards are presented in the same order in which they were presented to each subject. The actual length of each line and the hinted length follow: Real Hint 1. 4" 2" 2. 8" 10.5 3. 2" 4" 4. 8" 5" 5. 9" 13" 6. 5" 2.5" 93 9k 95 96 i ! • ; I 98 APPENDIX D The Five-Point Continuum Stimulus Card Used for the Post-Task Evaulation x I J y i x i * i ) w u c * i worse worse tfrnw the so.rue better -than Thuth better than most most other as most Thost other than *mcst other people people other people people other people APPENDIX E Comparison of the Means of the 42 Subjects Who Were New Intake Patients and the Total 48 Subjects Table 20 Mean Inches of Error in the Hinted Direction for A l l Conformity Items Group 42 Subjects 48 Subjects High Depressed 4.23 4.23 Low Depressed 6.18 6.31 Success Condition 4.46 4.56 Neutral Condition 4.80 4.85 Failure Condition 5.94 6.40 Males 4.98 4.90 Females 5.16 5.64 Table 21 Means of the Post-Task Evaluation Group 42 Subjects 48 Subjects High Depressed 3.37 3.37 Low Depressed 3.27 3.20 Success Condition 3.28 3.25 Neutral Condition 3.07 3.06 Failure Condition 3.64 3.56 Matles 3.36 3.33 Females 3.30 3.25 103 Table 22 Means of the Four Performance Predictions Group 42 Subjects 48 Subjects High Depressed 9.55 9.55 Low Depressed 10.50 10.30 Success Condition 9.23 9.40 Neutral Condition 9.69 9.57 Failure Condition 10.94 10.79 Males 10.29 10.23 Females 9.58 9.61 Self Prediction 9.71 9.70 Stranger Prediction 9.26 9.29 Mother Prediction 10.78 10.70 Father Prediction 10.07 10.00 Table 23 Mean Errors Made on the Matching Task Group 42 Subjects 48 Subjects High Depressed 3.95 3.95 Low Depressed 4.27 4.62 Success Condition 3.85 4.43 Neutral Condition 3.78 3.81 Failure Condition 4.64 4.62 Males 4.09 4.29 Females 4.10 4.29 Table 24 Mean Total Seconds Given as Time Estimates Group 42 Subjects 48 Subjects High Depressed 402.56 402.56 Low Depressed 431.00 475.75 Success Condition 401.85 498.68 Neutral Condition 425.96 413.96 Failure Condition 416.42 404.81 Males 351.61 358.68 Females 484.20 519.62 The actual total time for a l l items was 270 seconds.