GUTTMAN FACET ANALYSIS OF ATTITUDE - BEHAVIOR OF i ,_ PSYCHIATRIC PATIENTS AND NORMALS TOWARD THE MENTALLY ILL: CONTENT, STRUCTURE AND DETERMINANTS, . Thesis for the Degree'of Ph. D. MICHIGAN STATE UNIVERSITY ROSS MAYNARD WHITMAN 1970 III! III IIIMILIII III (I! IIIIIIIII III III III I W e_ ‘-- «~- L {B F: A R Y Michigan State University ...¢ This is to certify that the thesis entitled GUTTMAN FACET ANALYSIS OF ATTITUDE-BEHAVIOR OF PSYCHIATRIC PATIENTS AND NORMALS TOWARD THE MENTALLY ILL: CONTENT,‘ ' STRUCTURE AND DETERMINANTS presented by Ross Maynard Whitman has been accepted towards fulfillment of the requirements for _PLD_-degree inLQELLSfiling, Personnel Services & Educational Psychology I k/‘fi/ L/\ [/4147’2 //q/Z ’1'\ I r ’/7 Ma'o x’xv’lrofessor L/ll ] [El/p August 21, 1970 I)ate I 0-169 ABSTRACT GUTTMAN FACET ANALYSIS OF ATTITUDE-BEHAVIOR OF PSYCHIATRIC PATIENTS AND NORMALS TOWARD THE MENTALLY ILL: CONTENT, STRUCTURE AND DETERMINANTS BY Ross Maynard Whitman Problem Very little systematic attention has been given to how either psychiatric patients or "normals" view mentally ill persons or what impact they might have on one another. In spite of the importance of intergroup attitudes in predicting, understanding, and controlling human rela- tions, the interdependent nature of attitude formation and expression on the part of one psychiatric patient for another has been all but overlooked. The importance of non-psychiatric as well as psychiatric patients' attitudes insofar as they relate to social structures, conceptualiza- tions of mental illness, deve10pment of treatment programs, and the development of training and educational programs for mental health workers has also been overlooked. The purpose of the present research was to conduct a comprehensive investigation of attitudes toward mentally Ross Maynard Whitman ill persons, as well as to assess the predictive validity of certain hypothesized determinants of attitudes, with selected groups of psychiatric and non-psychiatric persons. Methodology The Attitude Behavior Scale-Mental Illness (ABS-MI) was administered to three samples of hospitalized psychi- atric patients, i.e., schiZOphrenics, alcoholics, and elderly persons, and to a sample of "normal" subjects. The ABS-MI was used to compare the attitudes of these samples to four classes of hypothesized determinants and/or predictors of attitudes: (a) contact, (b) knowledge, (c) values, and (d) demographic factors. Guttman's facet theory of attitude structure and his definition of attitude as "a delimited totality of behavior with respect to something," guided the construction of the ABS-MI. Facet theory specifies that the structure of an attitude universe is multidimensional and that it can be sub-structured into attitude levels which are system- atically related according to the number of identical con- ceptual elements they hold in common. The relationship among the various attitude levels is one in which the atti- tude levels closest to each other in the semantic scale of their definitions will also be closest statistically and the resulting matrix of attitude level inter-correlations will assume a simplex ordering. Ross Maynard Whitman The ABS-MI measures six levels of attitude interaction with the mentally ill. Each level contains one more strong element than its predecessor on an abstract-impersonal to concrete-personal action continuum. They were labeled (a) Societal Stereotype, (b) Societal Interactive Norm, (c) Personal Moral Evaluation, (d) Personal Hypothetical Action, (e) Personal Feelings, and (f) Personal Action. . Kuder-Richardson-type reliability estimates obtained for the ABS-MI were on a par with other attitude scales. Its ability to differentiate groups chosen on the basis of presumed differences in attitudes toward the mentally ill provided support for concurrent validity. Moreover, simplex approximations, predicted by facet theory, were obtained for each group of subjects which provided additional sup- port for the construct validity of the research instrument. Results Combining ABS-MI content and intensity scores en- hanced reliability estimates.. Reliabilities were generally highest on the more abstract-impersonal attitude levels for the psychiatric subjects while they were highest for the normals on the concrete-personal action continuum. The relationship between several independent variables and attitudes toward the mentally ill varied across samples and the six levels of the ABS-MI. Enjoyment of contact pro- vided the most consistent positive relationship with atti- tudes across samples and levels. In addition, some Ross Maynard Whitman independent variables, most notably education, change ori- entation, efficacy, and amount of contact, were negatively related to the subject's perceptions of others' attitudes toward the mentally ill and positively to perceptions of their own attitudes. The attitudes of the normal sample tended to be most positive, while those of the elderly subjects were the least positive. The pattern of the relationships between the indepen- dent variables and attitudes toward the mentally ill re- peatedly demonstrated the multidimensional nature of the ABS-MI attitude levels. The simplex approximations which were obtained, as predicted by facet theory, also sub- stantiated attitude multidimensionality. A number of limitations of the present study1 and recommendations for further research were presented. 1The present study is related to a larger international study of attitude toward physical, mental, and racial-ethnic differences under the direction of John E. Jordan, College .of Education, Michigan State University, East Lansing, Michi- gan 48823. GUTTMAN FACET ANALYSIS OF ATTITUDE-BEHAVIOR OF PSYCHIATRIC PATIENTS AND NORMALS TOWARD THE MENTALLY ILL: CONTENT, STRUCTURE AND DETERMINANTS BY Ross Maynard Whitman A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Counseling, Personnel Services and Educational Psychology College of Education 1970 / ta 7.27 G; Copyright by ROSS MAYNARD WHITMAN 1971 ACKNOWLEDGEMENTS I am indebted to a large number of individuals for their interest, support and assistance. Foremost is Dr. John E. Jordan who served as chairman of my doctoral committee and guided the research from its formative stages through its completion. I would also like to express my appreciation to Dr. Gregory A. Miller and to the remaining members of my committee, Drs. Harvey F. Clarizio and Thomas Gunnings for support and suggestions. I am grateful also to Dr. Stewart G. Armitage, Chief of Psychology Service, Battle Creek Veterans Administration Hospital, for graciously granting me time and independence to complete the research. I am particularly indebted to the patients at the Battle Creek V. A. Hospital and to the rehabilitation and guidance students as well as their instructors, Drs. James Costar and James Engelkes for their cooperation. Comple- tion of the research could not have been accomplished with- out them. Finally, words cannot fully express my heartfelt gratitude to my wife, Jo Anne, for her unfailing encourage- ment and for the many sacrifices she has patiently endured. ii Our young son, Jason Paul, also merits special recognition for both the pleasant diversions and the extra motivation he unknowingly provided throughout the research. iii TABLE OF CONTENTS ACKNOWLEDGEMENTS . . . . . . . . LIST OF TABLES. . . . . . . . . LIST OF FIGURES . . . . . . . . Chapter I. INTRODUCTION . . . . . . . Nature of the Problem . . . Statement of the Problem . . II. REVIEW OF RESEARCH AND THEORY . Schi20phrenics and Normals . Alcoholics. . . . . . . Lay Public. . . . . . Caretaking Groups . . . Alcoholics. . . . . . Geriatrics. . . . . . . Aged Persons . . . . . Young Persons. . . . . Psychiatric Patients . . Facet Theory and Methodology. Guttman Four-level Theory. Jordan's Six-level Adaptation Maierle's New Theory Formulation Related Projects. . . . iv Page ii viii 11 ll 15 16 18 21 22 23 25 27 28 28 33 37 41 Chapter Page III. INSTRUMENTATION AND VARIABLES . . . . . 43 Measurement . . . . . . . . . . 49 Intensity . . . . . . . . . . 49 Combining Content and Intensity Scores . . . . . . . . . . 50 Validity of Attitude Behavior Scale Type Instruments. . . . . . . 55 Reliability of Attitude Behavior Scale Type Instruments. . . . . 58 Personal Questionnaire (Independent Variables). . . . . . . . . . 60 Demographic Variables . . . . . . 61 Change Orientation . . . . . . . 61 Chronicity. . . . . . . . 61 Contact with the Attitude Object . . 62 Efficacy . . . . . . . . . 62 Knowledge About the Mentally Ill . . 63 Attitude Object . . . . . . . . . 65 IV. DESIGN AND ANALYSIS PROCEDURES . . . . . 67 samples 0 C O O O l O O O O O O 6 7 Pretesting of the Instrument. . . . . 71 Administration . . . . . . . . . 74 SchizOphrenics . . . . . . . . 75 Geriatrics. . . . . . . . . . 75 Alcoholics. . . . . . . . . . 76 Normals. . . . . . . . . . . 77 Research Hypotheses. . . . . . . . 77 Relating Attitudes and Demographic Variables . . . . . . . . . 79 Relating Attitudes and Religiosity. . 79 Relating Attitudes and Change Orientation . . . . . . . . 80 Relating Attitudes and Chronicity Orientation . . . . . . . . 80 Relating Attitudes and Contact . . . 80 Relating Attitudes and Efficacy. . . 80 Relating Attitudes and Knowledge . . 81 Relating Attitudes and Group Membership. . . . . . . . . 81 Relating Attitudes and Multidimensionality. . . . . . 81 V Chapter Analysis Procedures. Correlational Statistics . . . . Analysis of Variance . . . . . Simplex Approximation Test . . . V. ANALYSIS OF THE DATA . ABS-MI Reliability . Research Hypotheses. Relating Attitudes Variables . . Relating Attitudes Relating Attitudes Orientation . Relating Attitudes Orientation . Relating Attitudes Relating Attitudes Relating Attitudes Relating Attitudes Membership. . Relating Attitudes and Demographic and Religiosity. and Change and Chronicity and Contact . and Efficacy. and Knowledge and Group and Multidimensionality. . . . . VI. SUMMARY, DISCUSSION, AND Problem. . . . . Related Research. Instrumentation . Design and Analysis. Results. . . . Discussion of Results ABS-MI Reliability Relating Attitudes Variables . . Relating Attitudes Relating Attitudes Orientation . Relating Attitudes Orientation . Relating Attitudes Relating Attitudes Relating Attitudes Relating Attitudes Membership. . Relating Attitudes RECOMMENDATIONS. and Demographic and Religiosity. and Change and Chronicity and Contact . . and Efficacy. . and Knowledge . and Group and Multidimensionality. . . . . vi Page 81 81 82 83 86 86 89 89 99 101 107 112 123 125 127 130 135 135 137 140 142 143 149 149 150 154 154 155 156 157 158 159 160 Chapter Page Recommendations for Further Research . . 161 Criterion Instrumentation. . . . . 161 Independent Variable Instrumentation . 162 Analysis . . . . . . . . . . 164 Hypotheses. . . . . . . . . . 164 Administration . . . . . . . . 165 Generalizability. . . . . . . . 165 Theory . . . . . . . . . . . 156 REFERENCES 0 C O O O O O O O O O O C O 167 APPENDICES O O O O O O O O O O O O O O 175 A. ATTITUDE BEHAVIOR SCALE: ABS-MI . . . . 176 B. STATISTICAL MATERIAL . . . . . . . . 202 vii Table l. 10. 11. 12. LIST OF TABLES ABS-MI Hoyt reliabilities for content and combined content-intensity scores for psychiatric and normal samples . . . . ABS-MI-age correlations for psychiatric and normal samples. . . . . . . . . . ABS-MI-amount of education correlations for psychiatric and normal samples . . . . ABS-MI-length of hospitalization correlations for psychiatric and normal samples . . . ABS-MI-stated importance of religion corre- lations for psychiatric and normal samples 0 O O 0 O O O O O O O O ABS-MI-stated adherence to religion corre- lations for psychiatric and normal samples 0 I O O O O O O O O O O ABS-MI-change orientation partial and multi- ple correlations for psychiatric and normal samples. . . . . . . . . . ABS-MI-product-moment correlations with change orientation variables for psychiatric and normal samples . . . . ABS-MI-chronicity orientation partial and multiple correlations for psychiatric and normal samples . . . . . . . . ABS-MI-chronicity orientation correlations for psychiatric and normal samples . . . ABS-MI-intensity-amount of contact with mentally ill persons correlations for psychiatric and normal samples . . . . ABS-MI-contact partial and multiple corre- lations for psychiatric and normal samples . . . . . . . . . . . . viii Page 87 90 93 97 100 102 104 106 109 111 113 116 Table Page 13. ABS-MI-contact correlations for psychiatric and normal samples . . . . . . . . 120 14. ABS-MI-efficacy variable correlations for psychiatric and normal samples . . . . 124 15. ABS-MI-knowledge variable correlations for psychiatric and normal samples . . . . 126 16. ABS-MI-adjusted means, 3's, and multiple means test results for psychiatric and normal sample . . . . . . . . . . 128 17. gz's for obtained and empirically ordered matrices on six level ABS-MI for psychiatric and normal sample. . . . . 132 18. N's, means, and standard deviations for psychiatric and normal samples on 12 dependent and 18 independent variables. . 203 19. N's, means, and standard deviations for psychiatric and normal samples on 1 independent and 6 dependent variables-- content and intensity combined . . . . 204 20. Sample sizes, adjusted means, E's, and multiple means test for psychiatric and normal samples--efficacy and ABS-MI content and intensity combined . . . . 205 ix Figure l. 2. LIST OF FIGURES Basic facets used to analyze semantic structure of an attitude item. . . . . Level, profile composition, and labels for four types of attitude item . . . . . Basic facets used to determine conjoint struction of an attitude universe . . . Attitude level, profile compositions, and descriptive names for six level attitude universe. . . . . . . . . . . . A definitional system for joint struction of the Attitude Behavior Scale-MR . . . . A mapping sentence of the joint, lateral, and response mode struction facets used to structure the Attitude Behavior Scale-MR . Combined content-intensity scoring procedure for ABS-MI levels 1-6 . . . . . . . Combined content-intensity scoring procedure for Efficacy-Life Situations Scale . . . Page 30 31 33 35 36 38 52 54 CHAPTER I INTRODUCTION To date, agreement on the definition of the concept of attitude does not exist. The sense in which this general term will be used in the present research follows the gen- eral orientation provided by Guttman (1950, p. 51) who defines an attitude as a "delimited totality of behavior with respect to something." For example, the attitude of a person toward the mentally ill could be said to be the totality of acts that a person has performed with respect to the mentally ill. Guttman's behavioral definition is used herein as it is more operational and lends itself to a facet theory analysis of the content, structure, and determinants of attitudes. Nature of the Problem The importance of attitudinal research is such that at present it occupies a central position in social psy- chology. Practically every textbook related to social psychology contains sections devoted to attitudes and their measurement. According to Newcomb (1954), there are two general kinds of reasons for measuring attitudes. First, measurement contributes to our social psychological theo— retical understanding. We can test our hypotheses about the conditions under which attitudes arise, persist, and change if we have measurements which can be related to the hypothetical conditions. Secondly, there are many practical applications of attitude measurements. We may want to know, for example, whether the attitudes of an individual or of a group are changing; or we may want to know how one person or one group compares with another. Without measurements, we should have to rely on estimates, guesses, or indirect evidence. Other authors have discussed pertinent reasons for studying the attitudes of institutionalized groups such as hOSpitalized mental patients. One group of authors (Reznikoff, Brady, & Zeller, 1959) write that recently there has been a growing awareness that the behavior and clinical course of the hospitalized psychiatric patient depends greatly on the intramural social forces with which he interacts. Greater recognition of the hospital as a therapeutic force has led to many innovations in psychi- atric hospital treatment--ranging from the prescription of specific attitudes that personnel should entertain toward patients to meet their needs, to the designing and utilizing of the social structure of the hospital itself as the main therapeutic force to bear upon the course of the patient's disorder. The latter approach has given rise to the concept of the hospital as a therapeutic community. Guttman and Foa (1951) have recognized that there is an increasing awareness of the part played by intergroup attitudes in controlling human relations. Studies on atti- tudes towards minorities, toward the mentally ill, and the like, are made in an effort to understand and perhaps to indicate how to influence the interplay between groups. The above authors conclude that the possibility of influencing an intergroup situation depends on the one hand upon the existing attitudes, and on the other upon changing the intergroup attitudes in the desired direction. Cohen and Struening (1962) conclude that despite the manifest importance of the area, there has been little systematic research directed toward the finding of rela- tionships between attitudes toward the mentally ill and such variables as symptom reduction, successful rehabili- tation of former patients, hospital discharge rates, etc. The authors point out that research of this kind depends upon the adequate conception and objective measurement of attitudes toward mental illness and the mentally ill. Social psychology and allied disciplines employ a myriad of techniques for the measurement of attitudes, but by far the most widely used and most carefully designed and tested technique is the attitude scale. Even though a disprOportionate amount of energy has been spent in re- search with attitude scales, much effort has been wasted because of the lack of suitable instruments for the measure- ment of attitudes and behavior. The present research is principally concerned with the theoretical orientation and methods of attitude measure- ment developed by Guttman (1959, 1961, 1966). His most recent contributions in the area of attitude scaling are entitled facet design and nonmetric analysis. In these approaches, Guttman has moved from the unidimensional realm of scaling into the multidimensional realm. This type of approach is advantageous in that it allows the researcher to examine many variables and their interrelationships simultaneously. Further elaboration of Guttman's provocative theory and research, as well as that of Jordan who has been instru- mental in stimulating both additional research and theoreti- cal underpinnings of attitude measurement, will be pre- sented in Chapter II. A thorough assessment and evaluation of attitudes toward the mentally ill constitutes a worthwhile undertaking inasmuch as mental illness is a crucial problem confronting the people of the United States and others of the world. Prevailing attitudes toward the mentally ill affects the perspective they have of themselves and others as well as how they are viewed by other peOple. And in turn, the ex- tent and quality of interpersonal relationships are affected (Farina & Ring, 1965; Swanson & Spitzer, 1970) as are social structures (Caudill, Redlich, Gilmore, & Brody, 1952). An examination of the literature reveals that con- siderable discrepancy exists regarding how psychiatric patients and normals view persons who are mentally ill (Crumpton, Weinstein, Acker, & Annis, 1967; Giovannoni & Ullman, 1963; Manis, Houts, & Blake, 1963). Also, research in the area of attitudes related to the mentally ill has tended to consider the subject primarily in terms of how parents, lay public, and professional mental health workers view and/or influence those who are mentally ill--the re- verse has not been subjected to extensive examination. While this focus has helped to eXpand knowledge concerning attitudes toward the mentally ill of certain segments of the population, it has tended to ignore the existence of attitudes held by one of the mental patients' most intimate associates--psychiatric patients. Little systematic atten- tion has been given to how psychiatric patients View men- tally ill persons. The reality of the truly interdependent nature of attitude formation and eXpression on the part of one psychiatric patient for another has been, with an occasional exception (Caudill, §E_al., 1952), all but over- looked in the research literature. Consequently, the atti- tudes of the mentally ill are far from clear and are in need of research designed to illuminate their structure, content, determinants, and relationships. Despite the importance of psychiatric patient atti- ‘tudes in terms of interpersonal relationships, social structures, conceptualizations of mental illness, develop- ment of treatment programs, and the development of training and educational programs for mental health workers, very little systematic research has been directed toward identi- fying factors which are instrumental in the development of attitudes toward the mentally ill. Generally, it is not known what importance can be attributed for different attitudinal levels to: (a) the amount of contact a person has had with the mentally ill, (b) the value orientation of the person, (c) the amount of factual knowledge about mental illness he possesses and ‘(d) the demographic characteristics of the subject. In addition, the literature is barren of research that utilizes, a facetized design (Guttman, 1959) to measure and analyze psychiatric patient attitudes toward the mentally ill. Moreover, it has been found that most attitude re- search is inconclusive or contradictory with reference to predictor variables. The reason, it has been suggested (Jordan, 1968), might be attributed to the fact that various measures of attitudes such as attitude scales were composed of items seemingly derived from different structures; i.e., from different levels of Guttman's sub-universes (Guttman, 1959), for example. Lack of control over the various atti- tudinal levels being measured seems likely to continue to produce inconsistent, contradictory, and non-comparable findings in attitude research unless an appropriate measure of attitudes is developed and employed. The present research is one in a series, jointly de— signed by several investigators. Common usage was made of instrumentation and theoretical material, as well as tech- nical and analysis procedures. The authors, therefore, collaborated in many aspects although the data were differ- ent in each study (Erb, 1969; Gottlieb, 1970; Hamersma, 1969; Harrelson, 1970; Maierle, 1969; Morin, 1969) as well as certain design, procedural and analyses methods. The interpretations of the data in each study are those of the author. Statement of the Problem Of the studies referred to above, Maierle's is most pertinent to this one. It had the following major purposes: (a) to prOpose a reformulation of the Guttman (1955) and Jordan (1968) theories and to make a preliminary test of that reformulation; (b) to construct, according to the formulations of Guttman, Jordan, and new formulations pro- posed by Maierle, an attitude scale with the emotionally disturbed as attitude object. The results of Maierle's completed research supported his formulations as well as those of Guttman and Jordan. He concluded that Guttman facet design provides a workable approach to attitude-scale construction; particular exten— sions suggested within the study are useful interpretations of the Guttman-Jordan approach; and that additional research is needed. The primary emphasis of Maierle's research was meth- odological, however, and did not encompass substantive aspects. Also, undergraduate students enrolled in an intro- ductory psychology course or an introductory education course served as subjects. Left unanswered, when psychi- atric patients serve as subjects, are questions about the feasibility of his attitude scales; the validity of Guttman's contiguity hypothesis; and the predictive validity of cer- tain hypothesized determinants of attitudes toward the emo- tionally disturbed/mentally ill. Truly, as Maierle and the research literature suggests, a need exists for additional attitudinal research of the kind designed to promote know- ledge of both substantive and methodological aspects. Therefore, the major purposes of the present research are as follows: 1. To construct, according to the formulations of Guttman, Jordan, and new formulations proposed by Maierle, an attitude scale with the mentally ill acting as the atti- tude object. 2. To measure and determine predominate value orien- tations and attitudes toward the mentally ill of the fol- lowing interest groups: (a) SchiZOphrenics, (b) Geriatrics, (c) Alcoholics, and (d) Normals. 3. To assess, in each interest group listed above, the predictive validity of the following hypothesized de- terminants of attitudes toward the mentally ill: (a) Demographic, (b) Valuational, (c) Contactual, and (d) Knowledge. 4. To compare a group of normals and selected groups of hOSpitalized psychiatric patients (schiZOphrenics, geri- atrics, and alcoholics) on the basis of: (a) their atti- tudes toward the mentally ill, (b) demographic, valuational, contactual and knowledge variables, and (c) the relation~ ships between a and b. 5. To replicate the six-level adaptation by Jordan of Guttman's four levels or types of attitudes, using Guttman facet design and analysis on a population comprised of normals and hospitalized psychiatric patients. This procedure serves to test Guttman's hypothesis that, accord- ing to his principle of contiguity, the matrix of level- by-level correlations will approximate a simplex (Guttman, 1959, 1966) which will be maintained across a group of non— psychiatric persons as well as groups of psychiatric patients. 6. To determine the effects of order of attitude scale levels on simplex approximations and the correspon- dence between a hypothesized order of administration of scale levels and a "best" order for the following groups: Schizophrenics, geriatrics, alcoholics, and normals. Knowledge attained through fulfillment of the pur- poses stated above may ultimately permit greater under- standing and prediction of the kinds of attitudes, 10 experiences, and situations which promote relatively nega- tive or positive behaviors of individuals and groups toward the mentally ill. In addition, knowledge gained through measurement of attitudes and their concomitants hopefully will provide a more solid theoretical and scientific founda- tion upon which to build prudent, efficacious programs designed to modify and expand the range of attitudes and behaviors that individuals and groups exhibit toward men- tally ill persons. CHAPTER II REVIEW OF RESEARCH AND THEORY Nunnally (1961) investigated the opinions and atti- tudes that normal subjects held concerning mental health and mental illness. He found that "the information held by the public is not really 'bad' in the sense that the public is not grossly misinformed but the attitudes held by the public are as 'bad' as is generally suspected." Schizophrenics and Normals In order to determine whether mental patients them- selves shared in the general public's negative attitudes, Giovannoni & Ullman (1963) compared the opinions and atti— tudes of hospitalized male psychiatric patients concerning mental illness with the opinions and attitudes of normal subjects presented by Nunnally (1961). The investigators concluded that hOSpitalized mental patients were found to be no better informed than normals about mental health. The patient attitudes toward the mentally ill were found to be as extremely negative as normals' attitudes toward the same group of individuals. 11 12 Manis, Houts, & Blake (1963) sought to compare hospitalized psychiatric and a control group of medical and surgical nonpsychiatric patients in terms of their beliefs about mental illness. The results of the study indicated that at admission to the Ann Arbor Veterans Administration Hospital the psychiatric patients' beliefs about mental illness were no more deviant when compared to expert opinion (staff members in the fields of psychiatry, psychology and psychiatric social work) than were the be- liefs of the controls. It was noted, however, that very severely disturbed psychiatric patients tend to view mental illness in more moralistic terms than do normals. The results of another study (Crumpton, Weinstein, Acker, & Annis, 1967) indicated that on the whole, the mental patient was described by hospitalized psychiatric patients and normals in unfavorable terms as excitable, foolish, unsuccessful, unusual, slow, untimely, active, cruel, weak, and ugly. However, patients differed signi- ficantly from normals on ratings of ten out of twenty seman— tic differential scales, giving in every case a more favor- able rating. Also, ratings of "mental patient" were some- what more likely to resemble ratings of "sick person" and "dangerous person" when made by normals and to resemble ratings of "criminal" and "sinner" when made by patients. The above authors conclude that their results support previous findings (Crumpton & Wine, 1965) which reflected 13 a similar picture and which showed that there are distinct differences between normal and patient views of mental illness. The findings of the latter study by Crumpton and Wine suggests that normal adults see the typical mental patient as sick but moral, and as a peculiar, different sort of person to be both pitied and feared. On the other hand, schizophrenic patients consider the typical mental patient as immoral rather than sick, safe but inconsequen- tial. According to Crumpton, §£_al. (1967), the image of the mental patient as seen by both patients and normals "is to say the least unflattering." But they believe that an unflattering image does not reflect "the extreme nega- tivism and antagonism that other studies have shown." On the other hand, the findings of Manis, et_al. (1963) tend to be somewhat contrary to those of other authors. Thus, the status of attitudes toward the mentally ill and mental illness is left in doubt. The review of the following related research litera- ture concerns attitudes of psychiatric patients toward mental hospitals, psychiatrists, treatment, and the rela- tionship between attitudes toward mental illness and demo- graphic variables. The main purpose of a study by Gynther, Reznikoff, and Fishman (1963) was to compare attitudes of patients toward treatment, psychiatrists and hoSpitals. A secondary 14 purpose was to determine if attitudes were related to demo- graphic variables such as marital status and age. The major findings from the data were that most patients ex- press relatively favorable attitudes toward psychiatrists, hospitals, and treatment upon admission. The least favor- able, although still positive attitudes, appear to be toward treatment. In addition, the data indicated that married patients have more positive feelings about treatment, psy- chiatrists, and hospitals than do patients who are not married. Older patients, furthermore, tend to perceive the psychiatric milieu more favorably than younger patients. In contrast to the findings of the above investiga- tors, Brady, Zeller, and Reznikoff (1959) found that favor- ableness of attitude toward treatment, psychiatrists and mental hospitals is unrelated to age, educational level, occupation, religion, duration of illness, or history of previous hOSpitalization. Gynther and Brilliant (1964) assessed the relation- ships among attitudes toward mental illness and demographic variables such as sex, age, education, marital status, and admission status. The results of this research showed that attitudes toward mental illness were not related to either sex or admission status; however, they were signi- ficantly related to age, education, and marital status. The data also indicated that older, married and less edu- cated patients are more custodial in their orientation 15 toward mental illness than the younger, unmarried, and better educated patients. Other investigators (Clark & Binks, 1966) have found that low age and high education are related to a liberal attitude toward mental illness. Alcoholics Advances toward finding solutions to the problem of alcoholism and its treatment and prevention can be best achieved through an understanding of the attitudes held by society, including other alcoholics, toward the alcoholic and treatment practices with alcoholics, as well as under- standing the disorder itself and the way in which these elements influence one another. Blane (1966) contends that the attitudes toward the alcoholic are largely hos— tile; treatment practices are used where least effective and our understanding of the disorder is limited. Further, these factors exert an unsalutary effect on each other. The research literature concerning attitudes toward the alcoholic is limited in both number and scope. What exists can be divided into three primary categories: (a) lay public, (b) caretaker groups such as physicians, psy- chiatrists, psychologists, social workers, and various other hospital personnel, and (c) alcoholics themselves. 16 Lay Public Traditionally, alcoholism, like mental illness, has been negatively viewed by the general public. In recent years, however, increasing public awareness and toleration of emotional factors in human behavior have fostered the emergence of an increasingly enlightened attitude toward alcoholism. A number of public opinion surveys undertaken in recent years have shown an apparently high acceptance of alcoholics as persons with an illness (Blane, 1966). A study by Mulford and Miller (1964) addressed itself to a somewhat more intensive examination of attitudes to- ward alcoholics than had been evidenced in previous studies. Their findings, representing the attitudes of adults in Iowa, show that approximately one-quarter of the pOpulation unqualifiedly accepted the alcoholic as an individual, but that the remaining three-quarters View the alcoholic in moralistic terms, although they may simultaneously see him as ill. One of every three adults in Iowa defines the al- coholic in moralistic terms as either "weak-willed" or "morally weak." Evidently, the notion of the alcoholic as a "sick" person has made some headway, but for the most part peOple continue to cling to the old stereotypes even while partially accepting a more contemporary point of view. Additionally, it was found that those intimately associated with an alcoholic were only slightly more inclined to accept a purely medical view, but were significantly less inclined to accept a purely moralistic view. 17 In another context, Freed (1964) studied the Opinions expressed by psychiatric hospital personnel and college students toward people who are alcoholic, physically dis- abled, or mentally ill. He found that both groups were significantly more accepting of physical disability than of the other two illnesses. The students were slightly, and the hOSpital personnel were significantly more accepting of alcoholics than of mentally ill peOple. However, the mean scores of both groups on measures of expressed Opin- ions concerning alcoholics anth e mentally ill were still within the non-accepting range. Freed concludes that his data support reports in the literature of generally nega- tive attitudes toward the alcoholic. Attitudes toward alcohol of undergraduate college students were correlated with scores obtained from measures of social maturity and impulse expression (Strassburger & Strassburger, 1965). The data demonstrated relationships betwen alcohol attitudes and the two personality variables. In general, students who were favorably inclined toward the use of alcohol and to treatment (rather than punish- ment) of the alcoholic scored higher than those opposed to alcohol on measures of social maturity and impulse expres- sion. The studies cited in this section show two elements of the pOpular view of alcoholics and alcoholism. First, alcoholics are viewed as morally weak or weak-willed and 18 second, that attitudes toward them are generally negative. Possibly, as has been suggested (Blane, 1966), peOple re- gard the alcoholic, like the criminal and the mentally ill person, as dangerous and incurable. Caretaking Groups What of caretaking groups such as physicians, psy- chiatrists, social workers, and nurses? Do they share the popular view? Chafetz and Demone (1962) maintain that among all caretaking professions, an ignorant, moralistic, and puni- tive attitude is rampant and continues to thrive. The investigators state that society deSpises the alcoholic while the medical profession shuns him. Commenting on the attitudes of social workers, Sapir (1958) concludes that in general they tend to avoid involvement with alcoholics. They are viewed as either too difficult to work with or too hopeless to help. Such comments strongly suggest that members of care- taking groups by and large perceive alcoholics much as the general public does. Much of the research in this area has been directed toward physicians, but there have also been occasional studies of public health nurses, psychia- trists, and psychiatric hospital personnel. Until recently, studies of medical attitudes have utilized survey methods rather than attempts to scrutinize attitudes in depth. The findings of these studies are nevertheless in agreement on 19 one major point in that attitudes of physicians toward alco- holics are largely negative. One of the early studies of physicians' activities and attitudes was by Riley and Marden (1946) who interviewed a sample of physicains. Almost one-half considered their alcoholic patients as uncooperative and over one-quarter of the physicians considered them as a nuisance and un- manageable, as well as creating specific annoyances. Fur- ther evidence of negative feelings expressed by physicians is provided by Straus (1952). His work reveals attitudes on the part of physicians that range from frank rejection of the alcoholic patient to more covert attitudes of rejec- tion expressed as discouragement. Physicians in Straus's study indicated that they did not like to treat alcoholics and found them disgusting. Alcoholics were descirbed as having no motivation for help and unwilling to assume re- sponsibility. In addition, it was thought that they de- prive psychiatric patients of care by consuming the physi- cian's time in fruitless efforts. An outspoken picture of physicians' attitudes is painted by Abrams and McCourt (1964). These authors admit that they were not immune to some of the attitudes they attributed to physicians in emergency wards. Feelings of hopelessness, anger, and disgust were not uncommon in them- selves in dealing with homeless, jobless, and unc00perative alcoholics. 20 The attitudes of some psychiatrists parallel those of a number of physicians. Hayman (1956) reports that even among psychiatrists, negative feelings toward treating alcoholics can be found. The attitudes of caretaking groups other than physi- cians and psychiatrists have been infrequently investigated. One of the few efforts in this direction is a study of public health nurses undertaken by Blane and Hill (1964). A notable finding is that nurses are generally more accept- ing of alcoholics than physicians seem to be. Like physi- cians, nurses define and identify alcoholism in terms of severe physical, psychological, and social deterioration, and many nurses are discouraged about rehabilitation ef- forts. Nearly half, however, find work with alcoholic patients challenging and the results of their efforts re- warding. Examination of the literature on attitudes of the general public and of caretaking groups toward alcoholics and alcoholism reveals considerable negativism. Alcoholism is ambiguously viewed as an illness and as a moral afflic- tion while alcoholics are seen as ill, morally weak, dis- gusting, and irresponsible. No doubt these attitudes affect the behavior of the general public and caretakers in their interactions with the alcoholic, often to his detriment. 21 Alcoholics Attitudes of alcoholics themselves have been directed toward alcoholism, self concept, and rehabilitation. Mindlin (1964) measured the attitudes of three groups of non-psychotic, hOSpitalized alcoholics toward drinking and alcoholism as well as self-esteem. The three groups con- sisted of those who had previous experience with psycho- therapy, those who had previous eXposure to Alcoholics Anonymous, and those who had not previously sought help of any kind. The experimenter found that alcoholics in the third group had the poorest attitude toward treatment and the highest self-esteem of the three groups. Self- esteem was lowest in the group which had been exposed to psychotherapy. Attitudes of two alcoholic groups toward rehabilita- tion were compared by Mechanic (1961). The sample for this study consisted of patients consecutively admitted to two rehabilitation centers. One center was a voluntary hospital with Open doors and a general therapeutic milieu, while the other was a locked ward with no treatment or informa- tional program. Mechanic's findings suggest that a thera- peutic milieu has a favorable effect on attitudes of alco- holics toward rehabilitation. However, as a consequence of a likely selection differential of subjects for the study, the validity of the results is somewhat questionable. 22 In another study (Wolfensberger, 1958) a group of committed alcoholics and a group of non-alcoholic patients were compared on the basis of their attitudes toward the mental hospital. The findings from the study revealed that the mental patients as a group had much more critical atti- tudes toward the hospital than did the alcoholic patients. Not unexpectedly, alcoholic patients who later escaped from the hospital had more critical attitudes than those who remained. Age, education, and incidence of previous hOSpitalization were found to be unrelated to the attitudes of alcoholics. Geriatrics Psychiatric problems and attitudes of aged persons present complex considerations for many segments of society. A comprehensive discussion of them, particularly with refer- ence to the possibility of promoting mental health and pre- venting or mitigating mental illness requires that these problems and attitudes be viewed in their social and psy- chological contexts. Professional peOple of various persuasions are be- coming increasingly concerned with the mental health prob- lems and attitudes of the aged as a whole--those resident in the community and those who come to psychiatric atten- tion in hospitals or other treatment facilities. A par- 'ticu1ar problem is the planning and provision of facilities 23 and services for this important and growing segment of the papulation. The magnitude of the problem of mental illness in the aged is very great indeed. There were over 300,000 such patients aged 65 years or older in long-stay facili- ties in 1963. Almost one-half of these patients had been diagnosed as manifesting a functional psychosis. The other one-half were diagnosed as having brain syndromes (Simon, 1968). How older peOple view the mentally ill is of inter- est because their attitudes are likely to influence their interactions with those who are mentally ill, and likely Ialso to influence the individual's own adjustment as he joins their ranks. But despite the size and importance of mental ill- fuass in the aged, as well as current expressions of interest if! geriatric patients, there is a surprising lack of at- tennpts to evaluate attitudes of aged patients toward mental iJJIness or the mentally ill. However, attitudes of elderly iIndividuals, young people, and psychiatric patients toward the aged and aging have been examined in a variety of s tudies . ésneg Persons Many studies contained within the research literature irHiicate that there is a substantial acceptance of stereo- tYpes about the aged individual in American society and 24 that the older one becomes, the more stereotyped these atti- tudes tend to be. Tuckman and Lorge (1952) compared the attitudes of three groups of older age subjects: a group living in the community, a group living in the traditional type of insti- tution for the aged, and an intermediate group living in an apartment house specially designed to house elderly peOple. Data gained by the investigators supported the hypothesis that as individuals become less able to function independently in the community, they subscribe to a greater degree to the misconceptions and stereotypes about old age. The apartment house group subscribed more to these miscon- ceptions than the community group, while the institutional population subscribed more than the apartment house group. Other studies (Merrill & Gunter, 1969; Tuckman, Lorge, & Spooner, 1953) support the finding that older subjects tend toward acceptance of stereotypes about the aged. Merrill and Gunter made a study of attitudes toward older people and preference for age of roommate in 100 patients hospitalized in double rooms. Their results are as follows: (a) more of the patients over 65 years of age preferred roomates under 65, (b) persons 65 years of age and older had more stereotyped responses toward old people than did either young or middle-aged groups, and (c) those over 65 years of age with only a grade school education had more stereotypic attitudes towards old people than did those with either a high school or college background. 25 The relationship between the beliefs of parents and those of their children toward old people has been investi- gated (Tuckman, gt_al., 1953). The correspondence in atti- tudes between parents and children is far from perfect. Parents subscribe more to the beliefs, misconceptions, and stereotypes about old peOple than do their children. Carp (1967) studied the attitudes of old persons toward themselves and toward others. He found that a nega— tive attitude toward the self characterized the inscrutable and those who lacked self-possession. A positive attitude toward the self tended to occur among those who expressed their feelings Openly and those not easily put Off balance or embarrassed. An additional finding was that attitudes, either positive or negative, toward other people were well established and resistant to change, even when the social context of life was drastically altered. Young Persons Much of the empirical work conducted in the area of how young people view Old age and the processes Of aging points toward two general conclusions: (a) old age is regarded as a period of life that is markedly different from earlier years, and (b) the differences are seen as predominantly, but not entirely, negative. Tuckman and Lorge (1953) investigated the attitudes of a young adult group of graduate students toward Old age. The results indicate substantial acceptance of misconceptions 26 and stereotypes about Old peOple. Graduate students in— volved in this study look upon Old age as a period charac- terized by loneliness, resistance to change, and failing mental powers. Tuckman and Lorge speculate that these results are a reflection of the cultural expectations re- garding the activities, personality characteristics, and adjustment of Older people. Certainly, the data compiled in the study indicate that Old peOple are living in a social climate which is not conducive to feelings of adequacy, usefulness, security, and to good adjustment in their later years. Results similar to those of Tuckman and Lorge (1953) were Obtained by Kastenbaum and Durkee (1964). They found that although students did not express harsh sentiments toward elderly persons and did express some highly posi- tive sentiments, nevertheless, the general impression was that of a bias against the aged. According to the investi- gators, younger individuals, perceiving that Older persons resent them, attempt to avoid interpersonal contact and partially justify such avoidance by suggesting that Older individuals are really more interested in their families and are preoccupied with death. On a more positive note, they were regarded as wise and kind. The writers conclude that evidence exists which indicates a predominantly nega- tive appraisal of Older peOple as well as a tendency for young adults not to take the later years of their own lives into consideration. 27 Some investigators (Drake, 1957; Tuckman & Lorge, 1954, 1958) have found little or no relationship between contact with the elderly and stereotypic attitudes toward them. An exception is provided by Rosencranz (1969). He found that young people who had close grandparent contact judged aged individuals more favorably than those having little or no contact with grandparents. Likewise, young peOple who had meaningful associational contact with at least one Older person exhibited more favorable attitudes toward aged people. Psychiatric Patients One study (Feifel, 1954) in the research literature had as its focus: (a) to determine the general attitudes of the mentally ill toward aging; (b) to examine the rela- tionship between level of adjustment and attitudes toward aging; and (c) to compare these with the reported findings on normal subjects. Psychiatric, hOspitalized male patients served as subjects. Their mean age was approximately 35; average educational level completed was the eleventh grade; and average IQ level was slightly above 100. The findings of the study indicated that the psy- chiatric patients generally view old age in a gloomy man- ner. The degree of mental disturbance in the patients had little seeming effect on their overall gloomy attitude to- ward Old age. It may be, as the investigators conjecture, 28 that attitudes toward Old age are heavily determined by a widespread social attitude or ideology, rather than by idiosyncratic experiences. In conclusion, it now seems appropriate to briefly consider some points of comparison between the views of elderly people that are held by young, elderly, and psy- chiatric persons. Findings described in this section sug- gest that all three groups tend to have a rather negative appraisal Of Older peOple and that tne young tend not to think ahead to the later years Of their own lives. Fur- thermore, elderly people are regarded as being bound to the past, an orientation shared by very few young adults. Facet Theory and Methodology The following section Of the present chapter com- prises a summary of both Guttman's original formulations and subsequent adaptations prOposed by Jordan and Maierle. Also related projects are presented. Guttman Four-level Theory Guttman defined attitude as "a delimited totality of behavior with respect to something" (1950, p. 51). Within the limits of such a definition, both verbal re- sponses and overt behaviors can be construed as attitudes. If the particular reSponse to an attitude item constituted an expression of attitude, then individual items could be analyzed to see what makes any kind Of ordering possible. 29 One way to order responses is to examine the various seman- tic factors involved in a particular response to a parti- cular item. For Guttman, the crucial point in the measure- ment of attitudes was the possibility of ordering a series of specific responses to instrument items and from that, ordering to make predictions. In a reading of an article written by Bastide and van den Berghe (1957), Guttman (1959) distinguished three semantic factors, or "facets," involved in a particular attitude reSponse: the subject's behavior (belief or overt action), the referent (the subject's group or the subject himself), and the referent's intergroup behavior (compara- tive or interactive). Each facet contains more than one Option or "element." Guttman distinguished between what he called "weak" elements and "strong" elements. The first Of the two elements of each facet was designated as the weaker one. Any particular attitude item, then, was as strong as the number of strong elements which appeared. Guttman arranges the facets for a particular project in what he calls a facet definition. The definition then contains the various facets and their elements in such a way that it reads like a sentence. Such an arrangement is then entitled a "mapping sentence." According to the rationale Of Guttman, if an attitude item can be distinguished semantically in terms Of three facets, then an individual item could have none, one, two, 30 or three strong elements--a total Of four combinations. Although eight combinations Of elements or semantic pro- files are conceivable, i.e., (1) a1 b1 cl, (2) a1 b1 c2 . . . (8)'a2 b2 c2, Guttman indicated a logical reason for considering only four permutations of strong-weak elements. If elements are correctly ordered within facets and facets are correctly ordered with respect to each other, a seman- tic analysis Of attitude items according to n dichotomous facets will reveal n + 1 types of attitude items. Guttman called these types "levels." The levels have an inherent order where each level has one more strong element than the level immediately preceding, and one less strong element than the level immediately following. Figures 1 and 2 are adapted from Guttman (1959). Figure 1 contains the three facets originally identified by Guttman and the elements identified within those facets. Facets (A) (B) (C) Subject's Behavior Referent Referent's Intergroup BeHavior a1 belief bl subject's group cl comparative a2 overt action b2 subject himself c2 interactive Fig. l.--Basic facets used to analyze semantic struc- ture Of an attitude item. Figure 2 contains the four permutations of strong- weak elements which Guttman identified and the descriptive names which he attached to each of those permutations. The 31 more subscript "2" elements a profile contains, the greater the strength Of the attitude. Level Profile Descriptive Label 1 alblc1 Stereotype 2 alblc2 Norm 3 albzc2 Hypothetical Interaction 4 azbzc2 .Personal Interaction Fig. 2.--Leve1, profile composition, and labels for four types of attitude item. If items are written to correspond to each of the four levels, Guttman (1966) hypothesized that levels closest to each other should be more similar and thus should corre- late more highly with each other than more distant levels. Therefore, the levels having the largest number of common elements should be most closely related. Thus level 1 responses should be more similar to level 2 responses than to responses of any other level. Guttman called his hypoth- esis'the "principle of contiguity." In essence, it implies that if the structure Of certain items are close semanti- cally, they should also be close statistically. The hypothesized relationship of levels is ascer- tained statistically by what Guttman calls a "simplex." A simplex consists of a matrix Of level-by-level correla- tions in which the order of the correlations is determined but not the exact magnitudes. A simplex thus defined, 32 would have a distinctive appearance. It would exhibit the characteristics of: (a) ascending correlations starting from the zero point (where the two coordinates meet) to the end points of either axes, and (b) closer correlations between adjacent levels than correlations separated by several levels. According to Guttman, if attitude items are correctly written, the matrix Of level-by-level corre- lations should approximate a simplex. Guttman's hypotheses have been supported by his own research (Guttman, 1961) as well as that by Foa (1958, 1963) and Jordan (1968). What Guttman prOposes to achieve by facet design and analysis is to construct the content of a scale by a seman- tic, logical, a priori technique and to predict the order structure resulting from empirical data. Guttman's theory and methodology is the reverse Of what in reality factor analysis accomplishes. Factor analysis is designed to try to make sense out of what already has been done by a mathe- matical process of forming correlational clusters and then naming them, i.e., calling them factors. As Opposed to this approach, facet design, in essence, names the facets before one begins. Hypotheses can next be tested empiri- cally by determining whether certain Cartesian products occur with the usual frequency and order expected, as shown by the relations among the features in the matrix Of level- by-level correlations or simplex. Thus, according to Guttman facet theory, it is possible to establish an 33 ordering principle so that the attitude structure can be "ordered" with some explicit a priori semantic meaning rather than attempting to a postori evolve the meaning by some procedure such as factor analysis. Jordan's Six-level Adaptation Guttman's paradigm, i.e., facet design and analysis, for attitude item construction allows for three facets and hence four levels. Theorizing that there might be other pertinent facets, but accepting those that Guttman identi- fied as apprOpriate, Jordan (1968) expanded facet analysis for attitude items dealing with specified groups to include five facets and hence six levels. Figure 3 indicates the facets and their elements identified by Jordan. (A) (B) (C) (D) (E) Referent Referent Actor Actor's Domain of Behavior Intergroup, Actor's Behavior_ Behavior al others b1 belief cl others dl comparison e1 symbolic a2 self b overt c self d inter- e2 Opera- action 2 action tional Fig. 3.--Basic facets used to determine conjoint struction Of an attitude universe. The product Of the five/two-element/facets of Figure 3 yield 32 combinations of elements or profiles. Not all com- binations are logical because Of semantic considerations and the selection of the "best" set Of components from the 34 32 possible is partly a matter Of judgement. As shown in Figure 4, six Of these profiles were chosen as psychologi- cally relevant, potentially capable Of instrumentation, and possessing a Specific relationship between themselves-- a simplex one. As B qualifies A's behavior in Figure 3, so E quali- fies C's behavior. Frequently, but not necessarily, A and C are identical; In such cases, B and E must be "con- sistent,‘ i.e., some combinations seem illogical; BlEZ' It should be noted that sometimes the subject filling out the questionnaire is identical with either referent or actor or both, but not necessarily so: i.e., in level 1 and 2 referent and actor are identical, the subject is asked to report about them; in level 3 the subject is iden- tical with the referent, but not with the actor; in level 4, 5, 6, subject, referent, and actor are identical. Joint struction in Figure 3 is operationally de- fined as the ordered sets Of these five facets from low to high across all five facets simultaneously. The more sub- script "2" elements a set contains, the greater the "strength" Of the attitude. It should also be noted that not all combinations are logical. The selection of a "best" group Of sets is still partly a matter of judgment. Two continua run through the facets: other-self and verbal- action. Figure 4 shows levels of attitude strength, the ele- :ment composition of the selected profiles, and a descriptive 35 term for each profile or level. Each successive level changes on only one facet so that the profiles have a simplex ordering. Type-Level Facet Profile Descriptive Name l alblcldlel Societal Stereotype 2 alblcldze1 Societal Norm 3 azblcldzel Personal Moral Evaluation 4 azblczdzel Personal Hypothetical Action 5 azbzczdzel Personal Feelings 6 azbzczdze2 Personal Action Fig. 4.--Attitude level, profile compositions, and descriptive names for six level attitude universe. Jointl struction applied to Figure 4 refers to the differences between levels on facets A through E. It is that part of the semantic structure of attitude item con- struction which can be determined independently Of specific response situations or attitude Objects. Figure 5, as pre- sented by Harrelson (1970), incorporates the data presented in Figures 3 and 4 and shows how the semantic structure of the six attitudinal levels of Jordan's Attitude Behavior Scale-MR (ABS-MR) (1967) is specified by the element com- position'or facet profile of each level. The counterpart to joint struction, which specifies attitude level, is "lateral" struction. The latter deals ‘ ( 1Previously labeled conjoint and disjoint by Jordan 1968). 36 .soHpmonmpmm HmpcszOHmOm AOH>mcmm mesppra map co coHposepm choH ma» non Empmsm HmaoHpHcHemn a .m .mHm m .cmcnmumn on» cpHs omcHuom N O COHpo< Hmsomnmm InmucH AHHHmspomv HHHmCOHpmmeO mmmnmommmo 0 mo ma mm :H uHmmmE mocmHhmaxm m>mc H .omopmumu on» Apnoea xchpv m H mwcHHmmm Hmcompom csz OuomnmucH OHHHmOHpmchOHn Hombmommmo m mo Nb mm H soc: Auoommmv mocmHnmaxo H No Hm uOH>mcmm .oooAMpmp on» csz pomnmch AoHsozv H H m m m Hmouumnpoasm H0 H3 mm m n o m e s HmCOmnmm HHHOOHumchOHn H O>mHHon H No Hm COHumsHm>m.OOOAmpon on» 29H: pomnmch AcHsocmv H 4 H m m Hmaoz Hmcompmm I HO HO mm 6-3 0 m o m HHHmOHpOSDOOHn mnmnuo O>OHHOQ H .OOOLMuOA. m H Enoz HmumHoom on» csz Opompmch onHsonwv HHHOOH HmeHOmeo m H0 H2 Hm Inocuoamn mpmcpo m>mHHmn mpmcpo .OOOAMSOA . H H weapompmum HmumHoom HHHmucwE on» on oopmdsoo OHHHmOH HmeHOHmHo H Ho Hp Hm Inocuoamc mamnuo m>mHHmb mnmnuo mEmz O>Hpmeome mpcmEOpmum HOCOHpHchmO mHHmoam pmomm Hm>mq 37 with the content of the item and is dependent upon a speci- fic situation and attitude Object. It Specifies and dif- ferentiates the content of the items of the ABS-MR through five additional facets, F through J. Figure 6 adapted from Harrelson (1970) consists of a mapping sentence showing the five additional facets specifying item content, or lateral struction, as well as the relationship between joint and lateral struction on the ABS-MR. In addition, facets K and L in Figure 6 illustrate the "response mode struction" of the scales which is the degree of favorable- ness and intensity with which the subject responds to the items structured by facets A through J. The ABS-MR is an instrument designed to measure atti- tudes toward mentally retarded persons and to contain six hypothesized attitudinal levels (Jordan, 1968). It repre- sents what is believed to be the first attitude scale con- structed on a priori basis according to facet theory. Data Obtained through the use Of this instrument supports the hypotheses of Jordan and Guttman (Harrelson, 1970; Jordan, 1969; Morin, 1969). Maierle's New Theorijormulation Maierle (1969) noted that Jordan did not examine fully (a) his choice of Specific facet permutatons for each attitude level; (b) the effect of order Of administra- tion on relative size of correlations among levels; and (c) the effect Of invariant directionality in answer foils. 38 .COHpmcpmpmm prcszmHmom poH>wnOm OOSDHOO< on» manposppm 0» com: mpmomm COHposnum woos uncommon can .HmnmpwH .ucHOn on» no mocmpcmm mcHOOmE < .m .mHm cch mH O>HuHmOO mx m Hmn0H>mnonIHa0Hmzcn MH AHnHmcmchv EsHoOE NH Aocm Oocon>v Hanusmc Hx Hcszv HOHHHQMmHO Hanuom Hm O>Huoouum HH 30H HH m>Humwmc x QOOHocmc noustnupm n O>HpchOo H HuHmcmch mocmHm> Hm>mq pHmne mmue uHmpe Asv Axv va HHV oneOzmem moo: mmZOEmmm mcoHpmHmn ozonw Humocoomm mu HHHEHC can xom u :oHpmHmemH new opmuHoz .cpHmmc mu HcH poopmpmn acmEHOHQEO u HHHMDCOE m m 203838 me 23 5H,: mpHdmwo s 30H m :oHumwhomn u HnHchv do enscomb m: HoomMOMMMEH EsHooE mm chH>Ho>ch mcoHpmHmp Ozonm HLmEHLO mm Op pomdmon csz H: O H HV cch Hm mpHmnu Hmcomnma u mmmooum coHumsHm>m mocmppomEH mcoHpmsuHm thq Ame on Aev oneoamem qmcom m.p0po< noH>mcmm co :HmEoo (muonmnmucH w.nouo< nono< poH>mcmm accumumm accumuom a: 2: SO 2: 2: ZOHBoDmBm BZHOW 39 Moreover, he pointed out that several related problems therefore remained: (a) identification Of all possible facet permutations; (b) examination of effects from order of administration and from item directionality; and (c) application Of apprOpriate statistical tests. Maierle's study was designed to resolve the problems mentioned above. He analyzed the 32 permutations of five dichotomized facets and found that only 12 appeared seman- tically possible--Jordan's six and an additional six. These 12 level members were hypothesized to be ordered within seven semantic paths, or ordered sets of level- member items. Each path comprised six or four of the 12 level members. Next a set Of experimental instruments was developed which dealt with the emotionally disturbed rather than the mentally retarded. The instrument consisted Of two for each of the semantic paths (one instrument with the level members in a random order and one instrument with the level members in the hypothesized order). Items were written so that while content Of answer foils remained identical, directionality Of foils and grammatical emphasis within items varied randomly. Directions for items Of each level member were simplified and phrased in parallel form; all items were phrased so that all sets Of answer foils were identical. 40 The 14 varying scales were administered in random order, one to a subject, to 825 undergraduate students. Kaiser's Q? procedure (Kaiser, 1962), which evaluates indi- vidual correlation matrices and orders the variables with matrices, was used to analyze the data Obtained from the 14 sub-groups. A Q? value was determined for the hypoth- esized and best orderings of level members within each semantic path; for the random-order administrations of semantic paths, the Q2 value was also determined for the particular random order of administration. The results Of Maierle's study indicated that for six of the seven paths analyzed, the Q? value for the randomly administered, randomly ordered matrix was less than the Q2 value for the randomly administered, hypotheti- cally ordered matrix. In no case--either of random admini- stration or of hypothetically ordered administration-~did the hypothesized ordering of correlations generate the best simplex approximation. The hypothesized ordering principle (from no strong facets to all strong facets in succeeding level members) generally produced a better-than-random order, but never the best order. On the other hand, no general ordering principle which would improve on the hypothesized ordering principle was immediately Obvious. Many of the best orders appeared to involve few reversals from the hypothesized order. Order of administration was apparently not of pri- mary importance in simplex approximation. 41 Maierle concluded that the lack of an ordering prin- ciple Obviously better than the hypothesized one and the generally close correSpondence between hypothesized and best orders, suggested that the hypothesized ordering prin- ciple, the level members identified, and the orders hypoth- esized among those level members, are useful extensions Of the Guttman-Jordan formulations. Related Projects As noted in the Introduction, the present proposal is related to several studies. Jordan (1968) reported on an ll-nation study of attitudes toward education and physi- cal disability. Jordan's Attitude Behavior Scale-MR_ (ABS-MR) (1967) is the principal instrument in the studies of Gottlieb (1970), Harrelson (1970), and Morin (1969). Hamersma and Jordan's scale (1969), a refinement of the Jordan and Hamersma scale (Jordan and Hamersma, 1969), dealing with attitudes toward Negroes and Whites, has been developed from the work Of Jordan (1968) and Hamersma (1969). The original version of the scale also appears in the work of Erb (1969), who investigated racial prejudice and em- pathy. Maierle (1969) conducted a methodological study concerned with an application of Guttman facet analysis to attitude scale construction. The work Of Jordan and of Hamersma is of two types, attitude-scale construction and substantive research. The 42 work of Erb, Gottlieb, Harrelson, and Morin is primarily substantive research. Maierle's study is one of attitude- measurement theory. CHAPTER III INSTRUMENTATION AND VARIABLES This study employs a new instrument--the Attitude Behavior Scale-Mental Illness (ABS-MI) - (Maierle, Jordan, & Whitman, 1970). The construction of the ABS-MI was evolved by a Guttman facet theory rationale which makes it possible to construct items by a systematic a priori design instead of by the method of intuition or by the use of judges. Also, facet theory enables the specification of Object-subject relationships as well as situation con- tent in each attitude scale item. The six attitude levels of the ABS-MI (Appendix A) are: (a) Societal Stereotype; (b) Societal Norm; (c) Personal Moral Evaluation; (d) Personal Hypothetical Be- havior; (e) Personal Feelings; and (f) Personal Action. Each level member has one more strong element than the immediately preceding level member and one less strong ele- ment than the immediately following level members. NO element becomes weak once it has been changed from weak to strong (see Figures 4, 5, and 6--Chapter II). The six levels of the ABS-MI correSpond to the set of level members first identified by Jordan (1967) and 43 44 used in his original scale. Also, the levels correSpond to what Maierle (1969) designated as semantic path "C" in his research. Although other attitudinal levels and seman- tic paths exist, those which comprise the ABS-MI were selected for the following reasons: (a) they have been employed by researchers who developed scales to measure attitudes toward the mentally retarded (Gottlieb, 1970; Harrelson, 1970; Morin, 1969) and by researchers who mea- sured attitudes of Negroes and Whites toward each other (Erb, 1969; Hamersma, 1969); (b) the theoretical and struc- tural aspects Of the attitudinal levels and their rela- tionship to other semantic paths and levels, have been examined extensively in a methodological study by Maierle (1969) who found a lack of an ordering principle Obviously better than the hypothesized one and a generally close correspondence between hypothesized and best orders; (c) although selection of a "best" set of components from those possible is still partly a matter of judgment, from a clinical frame of reference, they appear to have relevance for both normal and psychiatric subjects and, therefore, seem particularly fruitful. All items in the six sub-scales of the ABS-MI evolved directly from the facet design illustrated in Figure 5 and correspond to the definitional statements for each atti— tudinal level. From the complete facet design illustrated in Figure 6, eight content items, each with a corresponding 45 measure of intensity (described in the following section), were selected for each of the six levels of the instrument so that the final attitude scale consisted of 96 items. In other words, each one of eight attitude items and a correSponding measure of intensity was repeated across all six levels or sub-scales with the items, only, being al- tered to conform to the facet structure Of the different levels. Consequently, the item content or "lateral struc- tion" was held constant so that the attitude structure or "joint struction" could be more easily assessed. The ABS-MI was developed to measure the attitudes of selected groups Of psychiatric patients and "normal" per- sons toward the mentally ill in the following eight scale areas: (a) Marriage-to; (b) Intelligence-Of; (c) Under- standing-Of; (d) Invitations-to; (e) Friends-with; (f) Eating-with; (g) Helpfulness-Of; and (h) Loans-to. These are areas in which peOple in general as well as psychiatric patients are likely to have strong, definite attitudes to- ward the mentally ill. Additionally, other researchers (Crumpton, et_al., 1967; Giovannoni & Ullman, 1963; Maierle, 1969) have included some of these areas in their research which dealt with attitudes toward the mentally ill. The choices within each set of answer foils are identical in phrasing--"agree," "uncertain," "disagree." Directionality of foils is dependent upon the phrasing of 46 each individual item. For computational purposes, "un- certain" is always scored "2" while the favorable response (either "agree" or "disagree," depending on item phrasing) is scored "3" and the unfavorable response is scored "1." The higher an individual scores within a given level or across levels, therefore, the more favorable or positive are his attitudes. AS indicated above, all items on all levels contained within the ABS-MI were written so that the choices within all foils were the same for all items. One Objection to such an approach to item formulation is that of the danger of response set or reSponse bias, or the tendency to answer all questions with identical foils in a similar fashion, independently of the content of a specific item. In an attempt to overcome the objection stated above, Maierle (1969) controlled, by randomization, the directionality and grammatical emphasis of the items included in his scales. All level members were presented in one Of four variant item phrasings. Selection of the Specific phrasing for each item was by random assignment. This procedure was employed to control for the effect of various types Of phrasing--in particular, the possible effect of response bias in an instrument where answer foils are uniform through- out. Maierle recognized that in spite of his procedures and innovations, the effects of response bias and analOgous pressures, such as social desirability, were not necessarily eliminated. 47 In the present study, the following procedures were incorporated in an effort to reduce the possible response bias of subjects: (a) each attitude item was followed by a question about intensity; (b) the order of the choices within answer foils were reversed or alternated on two levels of the scale; (c) someone was present to provide assistance or to answer questions while the subjects were completing the instrument; (d) all psychiatric patients were Observed while they were reSponding to the scale and, although they were allowed to finish, those who were Ob- viously responding in a set manner and/or not reading the items, had their data withheld from any further analysis; and (e) all subjects were given assurance that their re- Sponses were confidential and anonymous. Hopefully, these procedures Operated to reduce the number and extent of response sets. That they have been eliminated altogether is questionable, for attitude items such as those contained within the ABS-MI are verbalizations of behavior, or verbal expressions Of a set of behaviors. Such verbal attitude items are Open, to some degree, to the criticisms leveled at all questionnaire-type instruments. Questions concerning the effect of the order of scale level administration upon the resulting correlation matrix could be answered following completion of Maierle's (1969) research. In all of the previous research in this area, all of the data had been obtained from administration of 48 various level member subtests in the same order, i.e., all items of level 1 were presented first, all items of level 2 were presented second, and so forth. Maierle ad- ministered the scale levels Of a new Guttman type facet attitude scale in a randomly varied order and in a level— by-level order to a large group of subjects. He found that the matrix generated by listing correlations according to the hypothesized (level-by-level) order was a better Sim- plex approximation than that generated from the random order of administration. Maierle concluded that order of administration was apparently not of primary importance in simplex approximation and that he did not find an ordering principle obviously better than the hypothesized one, thus lending further support to the theoretical assump- tions involved. In view of Maierle's findings and conclusions, the ABS-MI was administered to all subjects in the hypothesized, or level-by-level order. The level names and choices of permutations of ele- ments, as well as the choice of item content, specific direction, and item formulations to match particular levels which have been included in this study, are not assumed to be the only possible ways in which such levels of attitude may be expressed. They are, however, consistent with the semantic implications of Guttman facet analysis. Therefore, the system of verbalizations that is presented is not 49 assumed to be all inclusive, but rather a convenient way to classify and relate variaties of behavior within a "delimited totality of behavior." Measurement Intensity Guttman and Foa (1951) have emphasized the importance of intensity measures in attitude scales, particularly with regard to the contact variable. Suchman (1950) has suggested that intensity of atti- tudes may be ascertained by asking a question about inten- sity immediately following a content question. This latter procedure was adopted by Jordan (1968) to measure intensity of attitudes on the ABS-MR. On levels l-5, the three a1- ternatives "not sure," "fairly sure," and "sure" are pre- sented to the question, "How sure are you of this answer?" for each item. A variation of this procedure was used on level 6 to ascertain whether a reported experience with the retarded was "unpleasant," "in between," or "pleasant." Both the Jordan system of facet analysis and the intensity dimension were employed in the construction of the ABS-MR Scale. The Scale has been administered to vary- ing groups of subjects and in most cases, the expected empirical simplex which was predicted from the contiguity hypothesis (Guttman, 1959) was approximated. Research by Maierle (1969) was primarily methodo- logical and did not concern intensity measures. As Guttman 50 and Foa (1951) point out, however, the use Of single ques- tions for the study of effect, change, or comparison is not advisable in that single questions ordinarily cannot dis- tinguish between changes due to intensity and those due to direction. In light of their observations, plus substantial research conducted by Jordan and his students (referred to in Chapter II), all of whom employed a measure of intensity and found it a useful addition, an intensity measure was used in the present research. Measurement of the intensity of attitudes on all levels of the ABS-MI was Similar to or identical to the procedure used by Jordan (1968) on levels 1-5 of the ABS-MR. ReSponses to questions regarding in- tensity of attitudes require a separate analysis and pro- vide a wealth of information not ascertainable by other means. Combining Content and Intensity Scores (a) ABS-MI.--In spite of the admonition of Guttman and Foa (1951) that it is inadvisable to ignore attitude in- tensity in comparing the attitudes of groups, Of those who have used both content and intensity measures to date, such as Maierle (1969), Morin (1969), Gottlieb (1970), and Harrelson (1970), only the latter paid heed to their warn- ing. The data reported in the research of the former in- vestigators referred to subject responses to either the content or intensity of each attitude item but not to the 51 two combined. Consequently, a weakly felt positive atti- tude response was given the same weight as a strongly felt positive reSponse. Accordingly, strongly and weakly felt neutral reSponses, as well as strongly and weakly felt negative responses, were weighted equally. Harrelson (1970) proposed that the content and in- tensity responses be combined into one score for each item to eliminate the seemingly inappropriate reSponse weights described above. Acting on his own prOposal, he develOped a rationale and scoring procedure for combining content and intensity reSponses which differentially weights each combination of these two variables for each item. This procedure is presented in Figure 7. The combined scoring procedure has the effect of increasing the range of possible scores for each item from 1-3 (negative to positive) to 1-9 (strongly negative to strongly positive). Theoretically, the reliability of each item should be enhanced, which in turn should further mag- nify group differences and relationships between the pre- dictor variables and the ABS-MI. The effectiveness of the combining procedure was tested and the results revealed that the ABS-MR reliabili- ties were elevated as predicted--particularly levels 1, 2, 3, and 5. Although the reliabilities increased on levels 4 and 6, their magnitude was of a lesser degree than the other ABS-MR levels. Additionally, the average item-to- scale level coefficients increased consequent to combining 52 SCORING PROCEDURE Content Alternatives 0--no response l--negative attitude 2--neutral attitude 3--positive attitude Intensity Alternatives 0--nO response l--weak intensity 2--medium intensity 3--strong intensity Combined Scores Content Intensity Combined 0 1 0 0 2 0 O 3 0 1 0 2 2 O 5 3 0 8 1 3 1 2 2 1 1 3 2 1 4 2 2 5 2 3 6 3 1 3 2 8 3 3 Rationale Delete from analysis because attitude direction is indeterminable Intensity error assumed and neutral intensity score of 2 assigned Strong negative attitude Medium negative attitude Weak negative attitude Weak neutral attitude Medium neutral attitude Strong neutral attitude Weak positive attitude Medium positive attitude Strong positive attitude Fig. 7.--Combined content-intensity scoring pro- cedure for ABS-MI levels 1-6. 53 content and intensity into one score. These results sup- port the ability Of this procedure to improve individual item reliability. In view of the advice rendered by Guttman and Foa (1951) and the pioneering research conducted by Harrelson (1970), and the results that he Obtained, it was decided to combine the content and intensity scores gathered from subjects who completed the ABS-MI. The results that are presented in this dissertation are based on data which con- sist of combined content and intensity scores unless other— wise specifically indicated. (b) Efficacy.--Prior research of Morin (1969), Gottlieb (1970), and Harrelson (1970), consisted, in part, of investigating the relationship between attitudes and a predictor variable labeled efficacy or life situations. Their data refers to subject responses to the cOntent Of each efficacy item and omits any reference to the intensity with which these items are responded to, although the effi- cacy scale (Appendix A) contains an intensity statement following each content item. Thus, a response to an effi- cacy item that was weak in intensity was given the same weight as a response that was strong in intensity and vice versa. Therefore, a scoring procedure and rationale was develOped for combining content and intensity responses which differentially weights each combination of these two variables for each efficacy item. This procedure is pre- sented in Figure 8. 5 SCORING Content Alternatives 1--Strong negative reSponse 2-—Negative response 3--Positive response 4--Strong positive response Combine 2222223 Intensity Combined 4 PROCEDURE Intensity Alternatives l--Minimum intensity 2--Mild intensity 3--Moderate intensity 4—-Maximum intensity d Scores Rationale l 4 1 1 3 2 1 2 3 1 1 4 2 4 5 2 3 6 2 2 7 2 1 8 3 1 9 3 2 10 3 3 11 3 4 12 4 1 13 4 2 14 4 3 15 4 4 16 Fig. 8.--Combined conten for Efficacy-Life Situations S Maximum strong negative reSponse Moderate strong negative response Mild strong negative response Minimum strong negative response Maximum negative response Moderate negative response Mild negative response Minimum negative response Minimum positive response Mild positive response Moderate positive response Maximum positive response Minimum strong positive response Mild strong positive response Moderate strong positive response Maximum strong positive response t-intensity scoring procedure cale. 55 The procedure in which content and intensity scores are combined has the effect of increasing the range Of pos- sible scores for each item from 1-4 to 1-16. As in the case involving the attitudinal items and levels of the ABS-MI, the wider range of scores should, if anything, in— crease the reliability Of each item and further differ- entiate group differences and relationships between the efficacy variable and those associated with the ABS-MI. Validity of Attitude Behavior Scale Type Instruments Many attitude studies are conducted for the stated purpose of systematically exploring verbally reported attitudes (Anastasi, 1961). In actual practice, what too Often happens is that investigators resort to a super- ficial kind of content validity based upon a cursory ex- amination and classification of the more stereotypic, com- parative, abstract and hypothetical aSpects of topics to Ibe covered. Most attitude scales do not contain content :related to verbalizations about affective experiences and (noncrete, overt behavior. The inclusion of level 5 (per- sonal feeling) and level 6 (personal action) in attitude behavior scale type instruments (Harrelson, 1970; Maierle, 11969; Morin, 1969) has the advantage of providing the Op- portunity to predict and analyze which of the other levels correlate highest with "personal feeling" and "personal ac tion " levels . 56 Guttman's facet theory (Guttman, 1959, 1961), which is fully described in Chapter II, specifies that the atti- tude universe represented by the item content can be sub- structured into components which are systematically related according to the number of identical conceptual elements they hold in common. The sub-structuring of an attitude universe into components or elements facilitates a sampling of items within each of the derived components, and also enables the prediction of relationships between various components of the attitude universe. The method of selecting item content on a system- atic basis through the use of facet theory and a mapping sentence, as was done in the case Of attitude behavior type instruments, appears far superior to other methods. It assures that a representative sample of the different be- havior domains is selected and that adequate content valid- ity of attitudinal levels is maintained. All items in the sub-scales of attitude behavior scale type instruments, as well as those included in the ABs-MI (Appendix A), evolved directly through use of facet analysis and a mapping sentence and correspond to specified definitional statements for each attitude level. Guttman (1959), analyzing the structure of attitude items, identified different levels, or types, generated from permutations of facets within each item. He hypoth- esized that, according to his principle of contiguity, the 57 matrix of level-by-level correlations would decrease in relation to the number of steps the levels were removed from one another and that they would approximate a simplex. Subsequently, Guttman's hypotheses have been supported by data Obtained by a number of independent researchers (Gottlieb, 1970; Harrelson, 1970; Jordan, 1970; Maierle, 1969; Morin, 1969), who have administered attitude behavior scale instruments to a variety of subject groups (univer- sity students, teachers, parents, and employers or manager executives) in a number of different countries (United States, Germany, Colombia, and British Honduras). Corre- lation matrices contained in the research of the investi- gators cited above, formed approximate simplexes as pre- dicted by Guttman's facet theory and his contiguity hypoth- esis. These results provide evidence of construct validity for the attitude behavior scales that were used and are generally supportive of facet theory as well as the utility and fruitfulness of the facet approach to attitude scale construction. Attitude scales can be validated sometimes through the use of contrasted groups as a special instance of con- current validation. Such an approach was taken by Jordan (1970), who chose subject groups on the basis of presumed differences in age, education, knowledge and experience regarding mental retardation, and cultural orientation. It was assumed that special education-rehabilitation 58 graduate students would have more positive attitudes toward the mentally retarded than education SOphomores and Belize (British Honduras) teachers, and that differences reflected on the attitude behavior scale instrument could be inter- preted as providing concurrent validation data. Analysis of variance results for the three sample groups revealed that the Special education-rehabilitation graduate students did in fact obtain a significantly higher total content score than either the education sophomores or the Belize teachers which provides some support for con- current validity of the attitude behavior scale that was used. Finally, data gathered by researchers (Gottlieb, 1970; Harrelson, 1970; Jordan, 1970; Morin, 1969) on a num- ber of predictor variables and the relationships between them and measures of attitudes obtained through the admini- stration of attitude behavior scale instruments Offer con- siderable "correlational" evidence of validity in that groups with known characteristics have tended to respond as predicted. Reliability of Attitude Behavior Scale Type Instruments The reliability of attitude behavior scales has been estimated by Obtaining a measure of internal consistency for each individual scale level by computing a Kuder- Richardson type reliability coefficient for each scale I 5. level (Harrelson, 1970; Jordan, 1970). In actuality, a variation of the Hoyt (1966) method, allowing for a dif- ference between the method of scoring the attitude behavior scales and the scoring method used in the Hoyt and Kuder and Richardson data was used to estimate the reliability of the scales. This method uses analysis of variance as described by Winer (1962) to provide a reliability coeffi- cient equivalent to the Kuder-Richardson measure of internal consistency. Harrelson, 1970, obtained combined score reliabili- ties which ranged from .60 to .92 and were consistently higher than the content-only reliabilities which ranged from .55 to .90. Average item-to-scale level correlations also increased through the combining procedure. The reli- ability estimates, and a pattern in which the lowest coef- ficients of reliability appeared on scale levels 1, 3, and 6 Obtained by Harrelson, were strikingly similar to those Of the groups used by Jordan, 1970. Thus, two separate researchers have used an attitude behavior scale instru- ment (ABS-MR) and have Obtained similar reliability coef- ficients and a similar pattern of these coefficients in the United States, Federal Republic of Germany, and British Honduras, and. in three different languages, i.e., English, German, and Spanish. These results suggest that various dimensions of attitudes in different countries were mea- sured with. approximately a similar degree Of reliability. 6O Reliability estimates on a large number of various types of attitude scales have been reported by Shaw and Wright (1967). The figures they present compare favorably with the majority reported for attitude behavior scales that have been developed in accordance with Guttman's facet theory and analysis. In addition, the reliabilities Of such instruments compare quite favorably to those of many tests used for individual diagnosis, evaluation, and selec- tion described by Anastasi (1961) . Therefore, the reli- ability of particular attitude behavior scales would cer- tainly appear adequate for research involving groups. Personal Questionnaire (Independent Variables) A comprehensive reivew of the literature (Jordan, 1968) on attitude studies indicated that four classes Of variables seem to be important determinants, correlates, and/or predictors of attitudes: (a) demographic factors such as age, sex, and income; (b) socio-psychological factors such as one's value orientation; (c) contact fac- tors such as amount, nature, perceived voluntariness, and enjoyment of the contact; and (d) the knowledge factor, i.e. , the amount of factual information one has about the attitude object. The instrument labeled Personal Questionnaire was designed to Operationalize several of the variables men- tioned above. Many of the items in this questionnaire were used in the international study of attitudes toward disabled 61 persons conducted by Jordan (1968) and subsequently by a mmmerofCMrdan's students (Gottlieb, 1970; Harrelson, 1970; Morin, 1969) . Also, it has been revised and adapted fertmelanrb (1969) and by Hamersma (1969) in their re- Inzwas further revised and adapted to meet the search. It is included needs and purposes of the present research. in Appendix A and consists of the various sections that follow. Demographic Variables Eight demographic items are included in the question- naire which from a theoretical standpoint might correlate with or predict, the criterion: age, amount of education, perceived importance of and adherence to religion, length‘ of hOSpitalization, marital status, religious preference, and type of major therapy experience. Change Orientation .Azmeasure Of change orientation was included to mea- self sure attitudes toward change in the following areas: change, child rearing practices, and birth control. Chronicity Ludwig and Farrelly (1966) maintain that the behavior of many psychiatric patients is governed by a code of The code is characterized by the patient's Chronicity. (b) lack desire for: (a) minimal involvement and thought, of reSponsibility for their own or other's actions, and 62 (c) greatest amount of privileges and fewest restrictions without corresponding Obligations. Three items were included in the questionnaire to measure Opinions with regard to the following areas: reSponsibility for self, involvement in activities, and the degree to which privileges should be earned. Contact with the Attitude Object Four items were designed to operationalize variables involved in personal contact between the respondents and the attitude Object. The items included are conceptually distinct. They report the following: the frequency of contact; the ease with which the contact might have been avoided; the amount of enjoyment experienced in the contact with the attitude Object; and the availability Of alter- natives tO associating with the mentally ill. Efficacy Attitude items which appear in the questionnaire under the heading "Life Situations" (Appendix A) were adapted from a Guttman Scale reported by Wolf (1967) . Measures Of intensity or answer "certainty" were added to the Original items. In addition, four levels of intensity of agreement- disagreement with the items replace the original "agree- disagree " dichotomy used by Wolf. The scale introduced above is designed to measure attitudes toward man and his environment and attempts to 63 determine the reSpondent's view of this relationship. This variable has been termed "Efficacy" by Jordan (1968) since the scale purports to measure attitudes towards man's effectiveness in the face of his natural environment. Knowledge About the Mentally Ill Nunnally (1957) constructed a 50 item Information Questionnaire which was administered to "laymen" groups consisting of 200 respondents in Knoxville, Tennessee, and 151 subjects in Eugene, Oregon. Also, the questionnaire was administered to an "expert" sample made up of 90 psychiatrists drawn from the Group for the Advancement of Psychiatry and 86 psychologists who were members Of the American Psychological Association. The results of Nunnally's study indicate that, for the laymen groups or general public, amount of knowledge about the mentally ill, and education, correlate positively Additionally, while age and knowledge correlate negatively. through the use of partial correlation analysis, it was found that younger persons and Older persons with the same number of years of education differ with regard to the amount Of knowledge possessed, and that the difference favors younger persons. The average responses by psychiatrists and psycholo- In no case was the difference gists were very similar. between the means of the two groups of any significance. The item variances for the experts were in all cases lower 64 than those for the general public and were less than one- half as large on the average. As indicated in Nunnally's data, the psychiatrists and psychologists tended to agree on the relatively simple and truncated kinds of information that the public at large deals with. Outside of the Older age groups and the low education group (more than 50 years of age and less than a high school education), the laymen responses were not markedly diver- gent from what the experts advocate in Nunnally's study. The public disagrees most with the experts on the same issues in which the experts disagreed most among themselves, regarding techniques for maintaining and restoring per- sonal adjustments. In the present research a 17 item knoweldge scale on mental illness (Appendix A) was extracted from the larger Information Questionnaire of Nunnally (1957). These 17 items were selected because they were specifically designed to measure the amount of factual knowledge possessed by the respondent regarding various aspects of mental illness. Also, the mean rating of each item by both experts and the general public were used as a basis of selecting the items. An item was accepted if both groups tended to rate the item (agree with the item) or below 3.5 (disagree) above 4.5 A number of the seven items on a seven point rating scale. with which the two groups agreed were given a mean rating of 5 or above while some of the ten items with which the groups disagreed were given a mean rating of 3 or below. 65 To offset possible response tendencies, the direction of each item was determined on the basis of a table of random numbers (Nunnally, 1957) . A table Of random numbers (Lindquist, 1953) also was used to determine the order of appearance of each of the 17 items on the knowledge scale. This procedure insured that items Of a true or false nature would be distributed throughout the scale on a random basis and would not be lumped together. All items on the information scale were in the form of declarative statements. The respondents indicated their agreement with each statement on a four point, Likert-type scale, i.e., strongly agree, agree, disagree, and strongly disagree. Attitude Object One characteristic of the present formulation is that the analyses involved are independent of a particular atti- tude Object, provided that object is some Specifiable group Of persons. The attitude object specified in the present studyis the "mentally ill" person. Such a person is described as manifesting "a disorder due to psychic causes, whether the symptoms are somatic, psychic, or behavioral" (English & English, 1958). Such a Specifica- tion rested in part on the assumption that the term "men- possessed a large enough commonality of meaning Two tally ill" as to be susceptible to measurement by verbal report. additional reasons for the choice of‘ this particular 66 attitude object was (a) the lack of research of the type proposed by Guttman and Jordan on attitudes toward the mentally ill, and (b) the lack of research regarding the attitudes of psychiatric patients, especially their atti- tudes toward other mentally ill persons. CHAPTER IV DESIGN AND ANALYSIS PROCEDURES The design of this research calls for the administra- tion of the Attitude Behavior Scale-Mental Illness (ABS-MI) and an accompanying Personal Questionnaire to samples of hospitalized psychiatric patients as well as to a group of subjects who were not hOSpitalized and "normal" psycholo- gically. Procedures designed to analyze the raw data were selected which would permit testing the relationships specified in the hypothesis. Samples The ABS-MI was administered to three samples of re- spondents at the Battle Creek Veterans Administration Hos- pital, Battle Creek, Michigan, and to a group of graduate students at Michigan State University. A description of the subject groups is presented below. (a) 117 Schizoghrenics.--A11 Of the subjects in this group were hOSpitalized Caucasian males who had been diag- nosed as having schizophrenia. They resided on open, un- locked wards located throughout the hospital. Their mean age was 41.42 years, with a range of 21—59 years and a standard deviation of 9.59 years. The educational level 67 68 of the group ranged from 4 to 19 years of school completed, with a mean of 11.32 years and a standard deviation of 2.45 years. The length of time spent in a psychiatric institution amounted to a mean of 9.02 years, with a stan- dard deviation of 7.21 years and a range of one month to 27 years. (b) 51 Geriatrics.--The psychiatric subjects in this group were comprised of hospitalized, elderly Caucasian male veterans, all of whom resided on open wards. They had a mean age of 68.78 years with a standard deviation of 7.42 years and a range of 60-91 years. One of the subjects had acquired only 4 years of schooling while another, a former dentist, had accumulated 21 years. As a group, the average educational level was 9.25 years, while the stan- dard deviation was 3.46 years. Psychiatric institutionali- zation ranged from a low of about one year to a high Of 46 years with a mean of 14.86 years and a standard deviation of 13.34 years. (c) 58 Alcoholics.--All of these subjects were Caucasian male veterans who had a history of using alco- holic beverages excessively and who, consequently, were housed on the alcoholic rehabilitation unit at the hospi— tal. They presented a mean age of 44.71 years with a standard deviation of 7.87 years and a range of 24—59 School attendance ranged from 6 to 17 years with a years. mean of 10.48 years and a standard deviation of 2.24 years. 69 When compared with the geriatric and the schiZOphrenic sub- jects as a whole, the alcoholic group had been hospitalized Length of hospitalization for a relatively short time. a standard deviation of amounted to a mean of 20.76 weeks; 32.65 weeks; and a range of 1-156 weeks. (d) 55 Normals.--The subjects in this group consisted of both male and female graduate students who were enrolled in either a course in principles of guidance and personnel services or a seminar in rehabilitation counseling. Due to the method Of data collection which was designed to facilitate computer analysis, exact (vs. coded) information with respect to the demographic variables of age, education, religious affiliation and marital status could not be Ob- However, in a general sense, the sample can be tained. under 30 years of age; graduate Of described as follows: a college or university; Protestant in religious orienta- tion; and either married or single. Originally, it had been planned to analyze the data separately for male and female students as well as for those Three students enrolled in the two different courses. First, factors led to the consolidation of the two samples. the number of male and female students in each class (par- ticularly the principles of guidance and personnel services class) was rather small and even their combined numbers in Second, for the each class could not be considered large. content scores as well as the combined content and intensity 70 scores for each level of the ABS-MI, no significant dif- ferences at the .05 level resulted when an F test was applied to test the significance of the ratio Of the two variances (Guilford, 1956, p. 224) for each of the following student groups: (a) combined male rehabilitation and guidance--combined female rehabilitation and guidance; (b) male rehabilitation-—female rehabilitation; (c) male (d) male rehabilitation--female guidance--female guidance; (e) male guidance; male guidance--female rehabilitation; rehabilitation--male guidance; and (f) female rehabilita- tion--female guidance. In addition, when the F test was applied to the scores on the change orientation items (101, 102, 103), Chronicity orientation items (105, 106, 107), and the contact items (108, 109, 110, 111) for each Of the above groups (a, b, c, d, e, and f), only one (item 108; male guidance--female guidance group) of a total of 70 dif- ferent F tests was significant at the .05 level. Third, the comparison of schizOphrenic, geriatric, and alcoholic patient attitudes toward the mentally ill with those Of normal subjects and vice versa, on a number of dimensions, constitutes one of the primary purposes of the present research endeavor. While a comparison of the attitudes of various groups Of subjects who are not hospi- talized in a psychiatric institution with one another it is beyond the pur- would be both of value and interest, poses and scope of this dissertation. 71 In view of the three points outlined above, the various groups of Michigan State graduate students were combined into one group Of normal subjects and their data were treated and analyzed as such in order to diminish the influence of chance factors and to increase the reliability and validity of the results. All of the points listed, especially the second, would seem to lend support and justi- fication to the procedure adopted. Pretesting of the Instrument It is essential that every new instrument be pre- tested before initiating the full-scale field operation Kahn & Cannel, 1957). Pretesting 1953; (a) to develop the procedures (Festinger & Katz, has at least three purposes: for applying the research instrument so that, for example, the instrument can be used effectively with respect to the time it takes to administer; (b) to test the wording of questions so that they are suited to the comprehension of the respondents; and (c) to ensure, as far as is practical, that the specific questions or Observations are really getting at the variable for which a measure is needed. According to the authors cited‘above, in the pre- testing of instruments and procedures, it is not essential to obtain a representative sample of subjects. It is im- however, to try to include some of the main types portant, There- of people who will be included in the final study. in the present study, the ABS-MI was pretested on 5 fore, schizoPhrenic and 5 geriatric subjects. 72 Pretesting revealed that the schizophrenic subjects generally were able to follow the instructions on the ABS-MI, understand the items, and complete the entire instrument in The fact that these subjects, a reasonable period of time. like most all Of those patients who eventually participated in the study, resided on unlocked Open wards, were not overtly hallucinating, and were in contact with reality, may account for their ability to comprehend and perform the task with a minimum of difficulty. Pretesting of the geriatric subjects disclosed a number of significant findings that had important implica- First, tions for the administration of the instrument. geriatric patients who had visual and auditory impairments and/or limitations of strength and movement in their upper extremities were generally unable to independently complete the instrument satisfactorily or at all. Second, the auditory comprehension of the subjects was superior to their visual comprehension. They tended to be slow and inaccurate readers and either did not comprehend the writ- How- ten material at all or misunderstood what they read. they demonstrated an ability to reSpond apprOpriately ever, directions and questions when they were read aloud to the Third, patients exhibiting moderate to by the examiner. manifestations of an organic brain syndrome were severe unable to perform adequately regardless of how the instru- Those who displayed mild disorienta- ment was administered. tion for time and place relationships and mild limitations 73 of their memory were able to complete the instrument ade- quately providing that the directions and items were pre- sented verbally and providing that sufficient interest and motivation could be maintained. Fourth, the geriatric patients questioned extensively and repeatedly the purpose of the instrument and the use to be made of their reSponses. They tended to request or demand assurance and support and required rather close supervision and guidance in order to resolve preoccupations with individual items and to main- :ain an orientation toward the task at hand. In addition, :heir interest in and motivation to complete the instru- ent appeared to depend to a large extent on the quality f the rapport that was established between the examiner 1d the patient. Pretesting revealed that the geriatric subjects rformed best when the instrument was administered indi- dually and when the instructions and items were read bud directly to them. Moreover, the Observations re- rted above, made it apparent that group administration the instrument to the geriatric subjects would be ill- ised, impractical, and unfeasible. Therefore, following testing Of the instrument, the decision was made to Lnister it individually to each geriatric subject al- tgh previously it had been thought possible to administer ABS-MI to them in groups. 1‘ 74 The instrument was not pretested on normal or alco- holic subjects. Since a number of the items had been administered to a sample of normal subjects by Maierle (1969) and in view of the results obtained from pretesting schizOphrenic subjects, it was not believed to be a neces— sary measure. Subsequent experience verified this decision. Administration The ABS—MI was designed for group administration, but it can also be administered individually. The instruc- tions and items are relatively simple and straightforward and the scale as a whole is comparable to the ABS-MR which requires about a fifth grade reading level. The instrument was administered to the schiZOphrenics in a testing room where extraneous noise and distractions It was given to the alcoholics could be held to a minimum. in an enclosed conference room that was located within the confines Of the building in which they reside. The geri- atric patients completed the instrument in the privacy of a quiet rOOm or an Office. Students at Michigan State University who were enrolled during the 1970 spring term in a course in either principles of guidance and personnel services or a seminar in rehabilitation placement, responded to the instrument as a part of their normal working class period. 75 Schiz0phrenics The ABS-MI was administered between January 16, 1970 and February 4, 1970 to a total of 126 Caucasian male schizophrenic patients who resided on Open unlocked wards. Various ward psychologists, nurses and nursing assistants believed that all of the 126 patients would be capable Of responding appropriately and adequately to the items on the ABS-MI. The performance Of only 9 subjects proved to be obviously invalid. accordance with a blatant position set; two were uncoopera- Two subjects answered the items in tive and responded to the items in a random fashion; two were unable to comprehend the instructions; two were overtly hallucinating; and four patients refused either to partici- pate at all or to continue to be involved in the testing. Consequently, the results obtained'from 117 subjects were used for the final statistical analyses. The average length of time required to complete the instrument was approximately 50 minutes and most of the schiZOphrenic patients finished it within an hour. Geriatrics The names of 73 Caucasian male patients who were 60 years of age or Older (born in 1909 or before) and who resided on an Open ward were Obtained from the medical Between December 9, 1969 and admini stration department . 1970 the ABS-MI was administered individually January 6 , The remainder to these subjects, of which 51 completed it. 76 were unable to respond apprOpriately to the scale because of mental confusion, lack of comprehension, or lack of adequate contact with reality. The length of time taken to complete the scale ranged from 45 to 90 minutes. Most of the geriatric patients took between 55 and 65 minutes to finish the scale while the average length of time re- quired was approximately one hour. Alcoholics Between January 5, 1970 and February 18, 1970 the ABS-MI was completed by 58 Caucasian male subjects who had a diagnosis indicative of excessive abuse of alcohol and rho had been admitted to the alcoholic rehabilitation unit t the hospital. Although a number of the patients had ne or more diagnoses denoting the presence Of a physical Eflication, only two had been diagnosed as having a :ychiatric disorder other than that related to alcoholism. e had a diagnosis of schizophrenia, latent type, while eacother's diagnosis was schiZOphrenia, chronic undif- rentiated type. All patients admitted to the alcoholic Tabilitation unit have free access to hospital ground vileges and they intermingle with one another as well as h other patients in both an incidental and planned hicn. The instrument was administered to the subjects in One .1 groups ranging from 3 to 8 members in number. ent was Openly hostile and negative about being tested 77 and, inasmuch as he purposely omitted data and was generally unCOOperative with respect to both behavior and attitude, his ABS-MI was invalidated and excluded from statistical analyses. One other potential subject could not be tested because he was not available due to his being on a pass, ill, etc., whenever a group of alcoholic subjects was tested. All but one of the alcoholic patients completed the entire instrument in one hour or less. A few finished in 30 minutes. The mean period of time needed to finish was approximately 45 minutes. Normals On April 30, 1970 the ABS-MI was completed by two rroups of graduate students at Michigan State University. ne group consisted of 21 male and 17 female rehabilitation ounseling students enrolled in a seminar in rehabilitation Lacement. One student did not finish the entire instru- Int and, consequently, was eliminated as a subject. The remaining group of students was comprised of 8 1e and 9 female students enrolled in a course in princi- zs Of guidance and personnel services. One student dis- tinued after responding to 36 items. Research Hypotheses A number of variables employed in this study were arcorrelated to enable examination of relationships for content and intensity scores of the criterion (ABS-MI) ‘7 I L..- .- 78 across each of six attitudinal levels with several inde- pendent variables. This procedure permitted testing of the research hypotheses. Research hypotheses that relate attitudes to variables concerning demographic factors (age and education), value orientations, contact factors, and a knowledge factor were derived from previous research (Jordan, 1968) related to the attitudes Of various groups in eleven nations toward the physically handicapped and toward education. In the present study, hypotheses were formulated to further expli- cate the relationship between attitudes of four subject groups toward the mentally ill and the four categories of predictor variables. Hypotheses pertaining to the relationship between attitudes and length Of hospitalization as well as atti- tudes and group membership stemmed from the work of prior & Reznikoff, 1969; Manis, Houts, researchers (Brady, Zeller, & Blake, 1963; Swanson & Spitzer, 1970) who found that persons hOSpitalized for a lengthy period of time held less favorable attitudes toward the mentally ill, psychi- atric treatment, psychiatrists, and mental hOSpitals than did persons hOSpitalized for a short time or those not hOSpitalized at all. The hypothesis relating attitudes toward the mentally ill and Chronicity orientation (see Chapter III) is based on the work reported by Ludwig and Farrelly (1966) who 79 Chlhmmtaithe characteristics of a code of Chronicity and its ramifications prevalent among psychiatric patients. The research hypothesis which deals with. attitudes andrmfliddimensionality stems directly from facet theory and Unaresearch of Guttman (1959, 1961), Jordan (1968), 1nd Maierle (1969) . Their research was discussed in thapter II. All of the thirteen research hypotheses contained ithin this study are presented below. elating Attitudes and gmographic VariableSI' h;i.--Persons who score BESS on age will score hgg I positive attitudes toward the mentally ill. H-2.--Amount of education will be positively related favorable attitudes toward the mentally ill. h:§.--Persons who have experienced a relatively art period Of hOSpitalization wil score higher on posi— re attitudes toward the mentally ill than will persons I have eXperienced a relatively long period of hospitalie ion: ating Attitudes and Religiosity ht:§.-—Persons who score high on stated importance Of Lgion will score 1931 on positive attitudes toward the :ally ill. I §h:§,-—Persons who score high on stated adherence to gion will score l_c_>_w_ on positive attitudes toward the ally 111. 80 Relating Attitudes and Change Orientation §;§u~Persons who score high on change orientation will score high on positive attitudes toward the mentally ill. gelating.Attitudes and :hronicity Orientation lhfl.—-Persons who score high on chronicity orienta- :ion will score high on positive attitudes toward the Ientally ill. Elating Attitudes and Contact h:§.--The more frequent the contact with mentally 11 persons, the higher will be the intensity scores on 1e ABS-MI, regardless of the direction (positiveness or egativeness) of attitude. H-9.--High freqpency of contact with mentally ill rsons will be associated with favorable attitudes toward e mentally ill i: high frequency is concurrent with (a) gernative rewarding Opportunities, (b) ease of avoidance the contact, and (c) enjoyment of the contact. rating Attitudes and Efficacy I+r10.--Persons who score high in efficacy will score _l'_1_ in positive attitudes toward the mentally ill. 81 Rglating Attitudes and Knowledge h:££.--Persons who score high_in knowledge about mental illness will score high in positive attitudes toward the mentally ill. gglatingAttitudes and groupMembership h;l3.--The groups will assume the following order Iith respect to favorable attitudes toward the mentally Ill: Normals>A1coholics>SchiZOphrenics>Geriatrics. elating Attitudes and ultidimensionaIity H-13.--The ABS-MI scale levels or attitude sub- niverses will form a Guttman simplex for each Of the ample groups. Analysis Procedures Control Data Corporation Computers (CDC 3600 and 6500) : Michigan State University were used to analyze the data. [grelational Statistics Considerable data can be employed in one analysis rough the use of the CDC 3600 MD-STAT program (Ruble & fter, 1966). Separate analysis can be done for the total cup and for any number of sub-groups, or partitionings the datai A number of statistics can be requested for :h specified group. Those used for each partitioning in is research.were means and standard deviations for each 82 variable and the matrix of simple correlations between all Variables. Partial and multiple correlations are outputs of the general multiple regression model used in the CDC 3600 program (Ruble, Kiel & Rafter, 1966a). The multiple cor- relation program yields the following statistics: the beta weights Of all predictor variables, a test of significance for each beta weight, and the partial correlations between each predictor and the criterion. One benefit of the use of multiple correlation is that a number of variables which are assumed to have some relationship to a criterion, or dependent variable, can be examined simultaneously. When a series of Pearsonian product-moment {'5 are computed between a criterion and a set of variables considered to be predictors Of the cri- terion, Spurious conclusions may be Obtained because the predictor variables are themselves inter-related rather than directly predictive of the criterion. In a partial correlation of each variable with the criterion, the effects of all but one variable are held constant. Analysis of Variance One-way analysis of variance statistics was calcu- lated by means of the UNEQl routine (Ruble, Kiel, & Rafter, 1966b) . This program is designed to deal with unequal frequencies that occur in the various categories. 83 A two-way analysis of variance design for unequal h's (Ruble, Paulson & Rafter, 1966) was used to analyze group-sex interaction which could occur only for the normal group and the total sample since they were the only ones that contained both males and females. Since none Of the samples were equal in size and some were not equal in sex ratio within groups (normal and total sample), all g_ tests were based on the adjusted means. Thus, the variance in the size Of the group samples was equalized or accounted for by the coefficients on which the adjusted means were based. Where it was applicable, the 3 statistic was used to test all of the mean differences. Several multiple means tests have been prOposed for determining the differences between treatment means (Winer, 1962) when three or more means are involved. In the pre- sent research, the 3 test for group comparisons was used to test for differences between adjusted means for pairs of groups equal to a two-tailed 5 test while also fully accounting for the other experimental factor. Simplex Approximation Test In Chapter II, a simplex was described as exhibiting 'the characteristics of descending absolute correlation cxoefficients moving from tOp to bottom in columns and aascending coefficients moving from left to right in rows. 84 Kaiser (1962) has formulated a procedure for scaling the variables of a Guttman simplex. His procedure orders the variables and suggests a measure of the goodness Of fit of the scale to the Obtained data. The approach developed by Kaiser may be seen as per— forming two functions: (a) a "sorting" of virtually all possible arrangements of data so as to generate the "best" empirically possible simplex approximation; and (b) an assignment of a descriptive statistic, "g2," to specified matrices. The index g2 is a descriptive one, with a range Of 0.00 to 1.00. A computer program has been develOped at Michigan State University which (a) re-orders the level member cor- relations Of each ABS-MI matrix by Kaiser's procedures, so as to generate the best empirically possible simplex approx- imation; and (b) calculates g2 for the hypothesized (theo- retical) ordering and for the empirically best ordering of each matrix. Kaiser's simplex approximation test has two rather serious limitations for which viable alternative procedures (m: not presently exist. First, the value 92 is a descrip- ‘tive statistic that has no test of significance available ‘bo evaluate how well a simplex has been approximated. Second, neither Kaiser's test nor the simplex model, which jg; based on Guttman's (1959) Contiguity Hypothesis, accounts for negative correlations. 85 As mentioned previously, there presently is no significance test available for Q2 values. The maximum number of order reversals that a 6 x 6 matrix could con- tain and still be accepted as approximating a simplex was set at six by Hamersma (1969). By this criterion, it was found that a 92 value of .60 was minimal and that to con- sider a correlation matrix as approximating a simplex, a value Of .70 should be used. As a result Of Hamersma's findings and recommendations, a Q? value Of .70 or above will be used to consider a matrix as approximating a simplex in this research. CHAPTER V ANALYSIS OF THE DATA The data were analyzed by computer at Michigan State University and the various analysis procedures used were described in Chapter III. ABS-MI Reliability The ABS-MI content and intensity scores were com- bined into one score for each subject on each item according to the procedure indicated in Chapter III. It may be re— called that the combining procedure resulted in increasing the range of scores possible for each item from 1-3 (nega- tive to positive) to 1-9 (strongly negative to strongly positive). The purpose of increasing the range of scores was to add to the discriminative ability as well as the reliability of each item and, thereby, effect an increase in the reliabilities of the various ABS-MI scale levels. The effects of combining ABS-MI content and inten- sity scores were tested by obtaining reliability estimates for the psychiatric and normal samples on content scores alone and on combined contents-intensity scores by the .Hoyt: (1966) method discussed in Chapter III. The reliability esthmates which resulted are presented in Table l. 86 87 mm. HS. mm. MS. So. am. coanEOO mHoEuoz mm Hm. So. em. HS. mm. om. acoucoo mm. mm. mS. mS. mm. Ho. cocHnEoo mOHHocOOH< mm mm. mm. 0S. OS. mS. mm. pcoucoo MS. MS. mm. mS. mm. mm. cocHQEOO mOHuuoHuoo Hm mS. mm. mm. mS. Hm. em. ucoucou em. mm. mS. HS. SS. SS. cocHQEOU mOHcoucmONHsom SHH we. em. SS. HS. oS. «S. usoucou o m a m N H mHSOHOHmmoou SHHHHQmHHom Ho>OH onom msouo .moneom Hoauoc can oHuuoHnommm How mmuoom SuHmcoucHIpcoucoo cOSHQEOO can ucoucoo How mOHuHHHnoHHOH whom HzImmmH oHoom HEImm< msouo .monEom HmEHOS can OHupmHnommm How mcoHHMHouuoo omMIHzImméII.m mqmme 91 support Hypothesis 1; however, it is not supported on normative level 2 or personal action level 6. Thus, com- pared tO younger persons, older persons in the total sample expressed more negative attitudes toward mentally ill per-’ sons on a stereotypic, moral evaluation, hypothetical action, and personal feeling level as predicted. They did not do so, however, on a normative or personal action level. Turning to the separate groups, one can determine that the statistically significant negative values Obtained on levels 3 and 4 for the schiZOphrenic and normal groups of subjects lend support to Hypothesis 1 for these groups, but it is not supported by the data shown for the other ABS-MI levels. On a moral evaluation level and a hypo- thetic action level, Older schizophrenics and normals ex- hibit more negative attitudes toward the mentally ill than do those who are younger. The results for the geriatric group indicate that‘ those who are older do not tend to manifest negative atti- tudes, as measured by the ABS-MI, any more than do those who are younger. The one exception occurs on stereotypic level 1 where a significant negative value of -.39 indi- cates that the more aged sample attributed more negative attitudes toward the mentally ill to others than did less aged subjects. 92 Resulting significant measures on levels 1 and 2 for the alcoholic group are in the converse direction of that predicted by Hypothesis 1. They suggest that alcoholics who are older attribute more favorable attitudes toward the mentally ill to other people, on levels 1 and 2, than do younger alcoholics. In summary, the data are largely supportive of Hypothesis 1 for the total sample, but less so for indi- vidual groups. It is supported only on ABS-MI levels 3-4 for both schizophrenics and normals, as well as only on level 1 for the geriatric group of subjects. None of the data for the alcoholics are supportive of the hypothesis, and the data for levels 1 and 2 are the opposite of what was predicted. h:§.-- Amount of education will be positively related to favorable attitudes toward the mentally ill. Hypothesis 2 was tested by correlating the ABS-MI with item 98 in the questionnaire which pertains to the Immount of education acquired. The resultant product- moment correlations, for the total sample and the four separate samples, are provided in Table 3. The measures in Table 3 reveal significant positive relationships between education and ABS-MI levels 3-6 for the total sample which is in direct support of Hypothesis 2. It is not supported for level 1 or for level 2 where a sig- :nificant negative figure of -.15 is in the direction 93 .Ho.v.m«« .mo.v.m« esoH. SNH. samH. «SNN.. asmH.I mo. mHmEmm HmuOB Hmm «mm. vo. smm. «mm. 00. Ho. mHmEHoz mm om. OH. OH. mo. So.I Ho.I mOHHO£OOH¢ mm «H. mo.I SH. So. vH.I mo. moHuuoHHow Hm oo.I Ho. mo.I Ho.I Ho. «HN.I moHeopeooNHeom SHH o m e m N H mcoHuoHouuou HO>OH OHcom HzImmm msouw .mOHmEom HoEuoc can OHuuoHsOSmd How mcoHuoHouuoo SOHHoOsco mo pcsoEoIHzImdeI.m mqmme 94 opposite to what was predicted. Comparison of the data for level 2 and 3-6, for the total sample, suggests that persons who have more education rate others as having an unfavorable attitude toward the mentally ill, while they assess their own attitudes as being favorable. Looking at the individual groups, it can be seen that there is a complete lack of a significant relation- Item-“'1' - .hh-I .4 Ship between attitudes and amount of education for the geriatric and alcoholic groups, as well as for levels 2-6 for the schiZOphrenics. The one exception for the latter occurs on stereotypic level 1 where a statistically sig- nificant figure of -.19 was obtained. Such a relationship suggests that the better educated schizophrenics consider other persons to have negative attitudes toward the men- tally ill and to compare them unfavorably. Significant positive correlations on moral evalua- tion level 3, hypothetical action level 4, and personal action level 6, for the normal group, comprise data which are in direct support of Hypothesis 2. Compared to the other groups (schizophrenic, geriatric and alcoholic), it indicates that the relationship between amount of educa- tion.and the ABS-MI is greater in the predicted direction on more attitudinal levels for the normal group. To summarize, the data in Table 3 support Hypoth- esis 2 on levels 3-6 for the subjects combined. It indi- cates that those who are better educated believe, more 95 Often than those less educated, they should (moral evalua- tion level 3), would (hypothetical action level 4), and do (personal action level 6) enter into positive relationships with mentally ill persons, and that they actually feel (personal feeling level 5) more positive about such rela- tionships. However, the results also indicate that, as a group, the better educated subjects see other people as thinking that "persons who are like themselves" (normative level 2) tend to hold more negative attitudes toward the mentally ill. h:§.--Persons who have experienced a relatively short period of hospitalization will score higher on positive attitudes toward the mentally ill than will persons who have experienced a relatively long period of hospitaliza- Hypothesis 3 was tested by correlating the ABS-MI with responses to item 104 in the questionnaire which asked the subject to report the length of his hospitalization in a mental hospital. It should be mentioned at this point that the arrangement of choices in the answer foil for item 104 is such that the highest numbered choice is asso- ciated with the shortest period of hospitalization. Con- sequently, positive correlations between the ABS-MI and item 104 indicate a positive relationship between attitudes and shortness of hospitalization, while negative correla- tions indicate the converse. 96 The correlation figures in Table 4 show that signi- ficant positive correlations on ABS-MI levels 3-4 and 6 resulted for the total sample, which is commensurate with Hypothesis 3. In other words, persons who have been hos- pitalized for a Shorter period of time are inclined to think that they or others should, they would, and they have or do associate positively with the mentally ill in a variety of circumstances. Interestingly, as indicated by a significant negative value on normative level 2, they do not think that other people have very positive attitudes toward the mentally ill. In addition to level 2, the data presented on levels 1 and 5 do not support the hypothesized relationship between attitudes and length of hospitalization for the combined samples. Looking at the groups separately, it can be seen that Hypothesis 3 is not confirmed by the data on any of the attitudinal levels for the geriatric and alcoholic groups. For these groups, there is not a strong relation- ship between duration Of hospitalization and favorable attitudes toward the mentally ill. Examination of the correlations given in Table 4 for the schizophrenics, reveals a strong positive relationship between attitudes and shortness of hospitalization on moral evaluation level 3 and hypothetical action level 4, which is supportive of Hypothesis 3. However, a significant :nelationship was not maintained for the remaining attitu- dinal levels. Therefore, it appears that schizophrenics 97 I, LIIIIIlIIl-i“ .Ho.v.m«s .mo.v.ms 4.0m. mo.- .NSH. AAHN. 44o~.I no.I oHosmm Hence Hem HH. asHv. same. ssmm. Ho.I No.I mHoEuoz mm mo.I So.I Ho.I eo. oo. oo. mOHHOLOOHN mm mH.I mH.I mN.I eN.I eN.I NH. mOHuuoHHoo Hm oH. we. 44mm. eeHm. Ho.I mo. moHeoneooNHeom SHH o m e m N H mcoHpoHoquO HO>OH oHoom HzImm4 msouo .monEom HmEuo: can OHHDoHsoame How mcoHHMHouHoo coHuoNHHonHmmos mo sumsoHIstmmdaI.e MHmOH OHoom HzImmm msouo .mOHmEom HoEHo: ccm OHuumHsomwm How mSOHumHouuoo conHHou mo oocmunomEH wmumwmlHEImmflll . m mqmflfi 101 Hypothesis 5 was tested by correlating the ABS-MI with item 100 in the questionnaire section which deals with how often one observes the rules and regulations of his religion. Examination of the results in Table 6 reveals a rather general lack of support for Hypothesis 5, with two exceptions. Statistically significant negative relation- ships occur On level 1 (stereotypic) for the schiZOphrenic group and for the total sample. On the other hand, sta- tistically significant (.05 level) positive correlations resulted on levels 3, 4, and 6 for the schizophrenics which is directly Opposite to the predicted relationship. None of the many remaining correlations between religious ad- herence and the ABS-MI were significant. It is apparent, as indicated by the results in Table 6, that the hypothesis is largely unsupported. Relatinngttitudes and Change Orientation h:§.-—Persons who score high on change orientation will score high on positive attitudes toward the mentally iii. A multiple correlation program was used to test Hypothesis 6. This program produced a multiple correla- tion coefficient between reSponses to three change orienta- tion items and each of six attitudinal levels of the ABS-MI for the total sample and each of the four sample groups .Ho.v.ms« 102 .mo.v.ae so. ~o.- Ho.I oo. oo. 4~H.I oHosom Hmnoe How oH. mm.I NN.I So.I mo.I HH.- mHesnoz mm oo. mo.I oH.I oo.I mo.I mo. moHHoeooHe om mo.I oo.I HH. mo. oH. mo. moHoueHeoo Hm 4mm. NH. aoH. 4mm. mo. 44mm.s moHeoneoqNHoom HHH o m o m m H -IIIIII- mcoHumHouuou HO>OH OHmom HzImmd msouo .mOHmEom HoEuoc can OHuuoHsommm HOM mcoHuoHouHoo COHoHHOH o» oocouocco coumumIHzImméII.o MHHNB 103 involved in the present research. The change orientation items pertained to self change (item 101), child rearing (item 102), and birth control (item 103). Results obtained from the multiple regression analy- ses are presented in Table 7. They lend general support to Hypothesis 6 for the total sample as well as for the geriatric, alcoholic, and normal groups of subjects, but not for the schizophrenic group. Specifically, multiple correlation data supports the hypothesis on levels 3-6 for the total sample; levels 2-5 for the geriatrics; levels 3-5 for the normals; and levels 2-6 for the alco- holics. The partial correlations presented in Table 7 are of particular interest since they suggest that after the effects of the remaining change orientation variables are "partialled out" or held constant, individual variables relate differentially to attitudes among subject groups, and that they make a differential contribution to the attainment Of a statistically significant multiple corre- lation coefficient. For example, three partial correla- tion coefficients achieved statistical significance on only the self change variable for the geriatrics and the total sample, but none of the coefficients for this vari- able reached significance at the .05 level for any Of the three remaining groups. On the other hand, three of the partial correlation coefficients were statistically "I‘- t L104 .Ho.v.m¢« .mo.v.o. .mH. ..oH. ..H~. ..om. ..-. so. eoHoeHonuoo oHoHuHoz oo. oo. oH. .NH. :.oH.I oo. Honosoo nonHm oHoeom Hobos How oH. so. oH. «HH. oo. No. ochmom oHHeo so. .HH. «HH. .HH. «NH. oo. mosseo eHom oH. .«om. ..ee. ..om. «Hm. HH. coHoeHouuoo oHaHuHoz oo. .4om. «.me. ..om. oo.- oo.- Hopscoo enuHm mHmenoz mm oo.- oo.- HH.- oo.- HH.- oo. ocHnmom oHHeo oH. oo. oo. so.I oH.I mH. mocmno oHom «om. «om. .mm. .om. .Hm. NH. aoHnoHonooo oHoHoHsz oo.- oo.- oo.- oH.I oo.- mo.- Houoaoo nuuHm moHHoeooHa om .SN. mm. NH. Hm. om. NH. ocHnmom oHHho no. mo. mm. om. oH. mo. mocmao eHom VH. saSm. same. item. «awe. mm. COHUMHOHHOU OHQHHHDE oH. oo. oH. mo. HH.- HH.- Hopscoo nonHm moHnooHuoo Hm oo. oo. om. om. oH. Ho. oeHumom oHHno oo.- .Hm. mm. .mm. «see. mo. ooomso eHom oH. oo. HH. mH. «Hm. oH. coHomHouuoo oHoHoHsz Ho.I vo.I oo.: oo. «ON.I So. Houucou :uuHm moHcounQONHnom SHH HH. No. oo. oH. so. oo.- ocHnmom oHHno HH. oo. oo. oo. oo. oo. potato oHom o m o m m H moHannm> mSOHHMHmuuou HO>OH OHoom HzImmd soHuoucmHuo woconu msouo .mOHmEom HoEuoc oso OHuumHgommd HON mcoHuoHouHOo deHuHSE can HmHuHMQ :oHumucoHuo moco£OIHzImmOH OHoom HzImm< msouu OmOOVOm. So. No. «oH. «NH. «NH.I «HH.- oHoeom Hmuoa HoN Ho. akSm. tame. «tSm. oo. So.l mHmEHOZ mm NH. No.I Ho.I mo.I No.I oo. moHHosooHe om Honueoo euuHm HH. NH. eN. HH. oo.- NN.I moHuuaHnmo Hm Ho. NH.I HH. oo.- Ho. oH.- moHeoueooNHoom SHH ««SH. oH. ««mH. ««SH. No. oo. oHosmm Hence HoN No. mo.- HH.- oo.- «SN.I oo. mHosuoz mm «SN. eN. NN. HN. eN. NH. moHHoaooHa om ocHummm oHHao oo. SH. «SN. «oN. NH. mo.- moHuumHumo Hm oH. so. So. oo. No. mo. moHcoueooNHnom SHH oH. «cmH. tamH. «aSH. SHH. mo.l OHmEmm HMUOB HmN SH. Ho.I oo.- oH.- HN.I SH. mHmsuoz mm oo. oo. oN. NN. oN. oo. moHHonooHa om mocooo HHom Ho.I ««mm. «Nm. ««em. ««me. No.I moHnumHuoo Hm oo. oH. So. oH. oo. mo.I moHconaooNHnom SHH o m o N N H noHnoHum> coHuoucmHuo mmcmsu .mOHmEmm Hmauos can OHHHMHAOSmQ How mOHQMHHm> coHpmucmHuo mmcmno suH3 chHHMHouuoo ucoEoEIuOSOOHmIHzImman.m mqmde 107 TO summarize, the results of multiple regression analyses tend to support Hypothesis 6 for all but the sample of schiZOphrenics. In addition, results repre- sented by partial correlation coefficients and product- moment coefficients indicate that the change orientation predictor variables relate differentially to attitudes toward the mentally ill and vary according to subject group and ABS-MI attitudinal level. The best sample group- predictor variable combinations are: (a) geriatrics-self change; (b) normals-birth control; and (c) alcoholics- child rearing. In view of the data, conceptualization of a single best predictor variable does not appear practical. Relating Attitudes and Chronicity Orientation h:1,--Persons who score high on chronicity orienta- tion will score high on positive attitudes toward the mentally ill. H-7 needs clarification to convey its prOper meaning and intent. Its present form is dictated by the answer foils contained within the ABS-MI and the Personal Question- naire and could easily lead one to think that persons who Inaintain a chronic orientation as described in Chapter III express positive attitudes toward the mentally ill. Actu— ally» those who receive a high score on questions concerning chronicity orientation, are low in terms of adOpting this frame Of reference. High scores indicate a rejection of a 108 chronicity orientation, i.e., lack of responsibility, dis- engagement from work, and a feeling that privileges should be granted automatically and not earned. Thus, a signifi- cant positive relationship between the criterion and the chronicity variable would indicate that positive attitudes are associated with low chronicity orientation. H-7 was tested by multiple correlation analyses which produced multiple and partial correlation coefficients be- tween the various ABS-MI levels and three chronicity orien- tation variables. The chronicity orientation questions in the Personal Questionnaire dealt with self responsibility (item 105), work involvement (item 106), and earning privi- leges (item 107). The results are presented in Table 9. The multiple correlation data in Table 9 generally are supportive of Hypothesis 7 for all but the alcoholic group of subjects. Thus, when the weighted composites of the three chronicity variables are considered, they serve as significant predictors of attitudes for most of the ABS-MI attitudinal levels for the total sample as well as for the schizophrenic, geriatric, and normal group of subjects. The partial correlation coefficients in Table 9 dis- close that when the effects of two of the chronicity vari- ables (responsibility and earned privileges) are nullified or held constant, the variable which deals with work involve- ment accounts for most of the variance within the largest 109 .Ho.v.m#t .mo.v.m« ««oH. ««oN. ««NN. s«oN. ««oH. ««mN. coHanonnoo oHoHuHsz mo. No.I oo.- mo. «NH. .«oH.I mooon>Hum oHosom Houoe HNN mH. ««NH. ««eN. ««oH. No.I oo. Homeo>Ho>oH xuoz Ho.I oo. So. «NH. oo.- ««oH.I SHHHHoncoomom oo. ««HN. ««Hm. ««oo. mN. «Hm. eoHnoHonuoo oHoHuHsz Ho. eo.I oo.- So. So. oN.I moooHH>Hum mHmenoz mm oo. oo. HH.- NN.I HN.- oo. neoso>Ho>sH Hues So.I HN. ««Hm. ««Se. oH. mo. NuHHHnHmsoomom NH. oN. HH. mN. «Hm. NN. eoHanonnoo oHoHuHoz eo. oH. oo. HN. HN. oH. moooHH>HuN moHHosooHe om oH. Ho.I oH. So. Ho. HH.- someo>Ho>cH Hues oo.- oo.- No.I No. mN.I oH.I SuHHHnHmcoomom NH. «om. «mm. «HN. NN. «.NN. eoHanonuoo oHoHuHsz No.I Ho. oH.- oo.- oo. mo. moooHH>Hnm moHuueHuoo Hm HH. oN. ««om. «om. NH. oo.- someo>Ho>eH xnoz oo. No.I oo.- HH.- oH.- «HN.- NHHHHonsoomom ««oN. ««SN. ««SN. ««mm. NH. «HN. ooHuoHonuoo oHoHuHsz mo.I oo. Ho. vH. NH. mH.I mommHH>Hum mOHcmHnmoNHnom SHH ««mN. ««oN. ««em. SH. oo. Ho.I HomeosHo>cH Hues oo. oo. oo. «oH. oo.- «NH.» SuHHHoncoomom o m o m N H MCOflHMHOHHOU H0>0~H OHMUm Hzlmmd moHnoHuo> SuHOHcOHso msouu .umHeEmm Hmfiuoc can OHuuanoxmd Mom chHumHmuuoo meHuHSS can HmHuuod coHuoucoHuo SuHOHSOHSOIHzIwmdtI.m mqmde 110 number of statistically significant multiple correlations for the schizOphrenic, geriatric, and total samples. But the variable which deals with responsibility contributes most to the variance in the multiple correlations for the group of normal subjects. None Of the partial correla- tions and only one multiple correlation (level 2) is sig- nificant at the .05 level, for the alcoholic subjects, which suggests that the chronicity orientation variables under consideration are not powerful predictors of atti- tudes toward the mentally ill for this group. Product-moment correlations which appear in Table 10 tend to mirror the results presented for the partial cor- relation data. Work involvement is the most powerful and significant predictor of attitudes for schiZOphrenics, geriatrics and the total sample, while self responsibility relates more powerfully to the criterion for the normal subjects than does the remaining chronicity orientation ‘variables. None of the chronicity variables relate sig- nificantly to attitudes toward the mentally ill for the alcoholics. In summary, the combined chronicity orientation variables relate significantly to ABS-MI levels 2-6 for the total sample; levels 3-6 for the schizophrenics, levels 3-5 for the geriatrics; and levels 3-5 for the normals in a way that is supportive of Hypothesis 7. The single most POWerful predictor of attitudes is work involvement for .Ho.v.mca 111 .mo.v.o« Ho. No. No. So. «NH. oo.- mHoSNm Hmuoe HoN No.I oo. oo. mH. Ho. «HN.- mHasnoz mm So. oH. oH. HN. HN. mo. moHHosooHa om moooHHsHuo mo. NH. oo. oo. So. No.- moHuuoHuoo Hm Ho. oo. oo. oH. oH. mo. moHconaooNHoom SHH «NH. ««oN. ««oN. ««NN. No.I oo.- oHoemm Hence HoN oo.- oN. oo. oo. SH.I oo.- mHmsnoz mm oH. oo. NH. NH. Ho.I mH.I moHHoeooHa om Homeo>Ho>eH Hues «H. «on. ««om. «oN. NH. HN.- moHnuoHuoo Hm oo. ««oN. ««oH. ««He. No. Ho.I mOHeoueooNHsom SHH Ho. «MH. SSSH. cewH. vo.l «NH.I OHmfimm HmuOB How oo.- «om. ««oe. ««eo. So. oo.- oneuoz mm oo. So.I no. mo. oH.- SH.I moHHosooHe om NHHHHoneoomom oH. NH. oH. No. HH.- ««NN.I moHuumHuoo Hm oo.- oH. NH. oH. HH. oo.- moHcougooNHnom SHH o m o N N H chHuoHouuou HO>OH OHnOm Hzlmma msouu annoHum> SuHOHcOHso .mOHmEom Hmeuoc can OHHHMHSOth mom chHumHmHHOO coHumucOHuo SuHOHcouno HzImm®H mflmom HElmm/N QSOHO .monEom HoEHoc can OHHuoHsommm HON mcoHuoHouuoo moomnoo HHH NHHmnaoe SHH; Homoeoo Ho oesosmISuHmooch HzImmeuI.HH mamas 114 do not support the hypothesis on any of the attitudinal levels for any of the other groups when considered sepa- rately. In conclusion, it would seem that for the subject groups as a whole, the more frequent their contact with the mentally ill, the higher is their intensity of atti- tudes toward them. Also, this relationship tends to be maintained for the alcoholic group if their results are considered separately, but it does not hold for the other individual groups. h:2.--High frequency of contact with mentally ill persons will be associated with favorable attitudes toward the mentally ill if high frequency is concurrent with (a) alternative rewarding Opportunitiesy (b) ease of avoidance of the contact, and (c) enjoyment of the contact. Originally, it had been planned to test Hypothesis 9 by means of a multiple correlation analysis which corre- lates the ABS-MI scale with the weighted composites of responses to items 108 (frequency), 109 (avoidance), 110 (enjoyment), and 111 (alternative), in the questionnaire section of the instrument. The product of a multiple cor— relation analysis, the coefficient Of multiple correlation, indicates the strength of the relationship between the ABS-MI and the four combined contact variables which served as partial correlates. However, full completion of the original plans was precluded because Hypothesis 9 requires 115 that statistical significance of the partial correlation coefficients of the alternative, avoidance, and enjoyment variables, for any one attitudinal level, must be achieved in order for it to be supported or unsupported, which are conditions that the statistical data did not meet. Examination of the coefficients of multiple and par- tial correlations presented in Table 12 discloses a general lack of concurrency of statistical significance among the contact variables on most all attitudinal levels for all subject groups. In fact, a single example in which all four contact variables were concurrent in terms of achiev- ing significance failed to occur. Moreover, except for the occurrence Of concurrency of three contact variables on level 6, for the geriatric group, no more than two contact variables were concurrent for any attitudinal level for any group. Thus, since the necessary variables were not statistically significant concurrently as specified by Hypothesis 9, it could not be tested as planned. Further inSpection of Table 12 reveals at least two important results which, although not a direct test of Hypothesis 9, seem to contribute to increased understanding and knowledge of the relationship between attitudes toward the mentally ill and specific contact variables. First, the multiple correlations generally are statistically sig- nificant for most of the attitudinal levels for all groups, indicating that a rather strong relationship exists between 116 .Ho.v.m«« .mo.v.m« «on. «oo. «oo. «mo. «mH. «NH. coHumHouuoo oHdHuHss oo. oo. oo. oo. No.I oo.- mo>HumcuouH¢ ««oN. ««mm. ««om. ««No. ««oH. oo. noossoHom oHosmm Hence HoN oo. «NH. HH. «NH. No. oo. oooooHo>< ««SN. HH. mo. Ho. oo.: oo.: ussosa «oo. «Ne. «HN. «mm. oN. NN. coHomHouuoo oHoHuHsz SH. SH. oH. oH. No. oH.: mo>HuoououH¢ NH. «HN. mo.I oo. mo.I So. nooesoncm mHMSHoz mm HH. oH. «oN. ««mm. No. No. oocooHo>¢ «HN. oo. mH. NH. oH. mo. ucoosa atom. «awe. same. «smv. «amm. «svv.. COHHMHOHHOU mdmwquz oo. No. oH. No. oo. NH. mo>HuocuouH< HN. ««SH. ««No. ««oo. ««Ho. Ho. ocossoncm moHHoeooHa om SH. NH. oH. oo. NH. mo.I mosooHo>< NN. Ho. oo.I oH.I ««NN.I ««Ho.I unsos< «tom. aaHm. «aHm. «SHQ. some. oN. COHUOHOHHOU OHQHHHSZ «NH. NH. No.I Ho.I So. SH.I mo>HuocuooH< «Hm. ««So. ««Sm. ««om. ««No. oo.- ucosaonom mOHuuoHuoo Hm So. So.I mN.I oN.- NH.I NH. oosmoHo>< ««mm. HH. oo. Ho. eo.I oH. ussos< ««SN. ««oN. ««He. ««oe. ««oN. mo. coHuoHouuoo oHoHuHoz oo. oH. oo. So. No. oo.- mo>HumcuouH¢ «oH. ««oN. ««Nm. ««om. So. Ho. namesoflom moHoonnooNHnom SHH oo. ««oN. ««NN. «HN. oH. No.I oooooHo>< SH. HH. NH. NH. ««oN.I Ho.I ossosa o m o N N H mCOflHMHOHHOU H0>U1H OHMUW HSImmC mmHnoHum> pumucoo maouu .mOHmEnm Hmfinoc can OHHHMHSOSmQ How mSOHuMHOHHOO meHuHsE can HoHuuom uoouSOOIHzImman.NH mamas 117 attitudes toward the mentally ill and various combined aspects of contact with them. Exceptions include the stereotypic attitudinal level for the normal, schiZOphrenic, and geriatric groups as well as normative level 2 for the normal group. Second, when the effects of the alternative, avoidance, and frequency contact variables are nullified or held constant, enjoyment of contact with mentally ill persons is generally strongly related to favorable atti- tudes toward the mentally ill on all levels except stereo- typic level 1, and for all grOups except that designated as normal. Notable exceptions to this second observation include level 2, for the schiZOphrenic group, and level 6, for the group of alcoholics which Obtained statistically insignificant partial correlations with enjoyment of con- tact. Also, none of the partial correlations between enjoy- ment of contact and the six ABS-MI attitudinal levels is significant for the normal group of subjects. Such a re- sult is contrary to that for the other three groups for which a large majority of the partial correlations for enjoyment of contact and the ABS-MI scales are significant. When the effects of the other three contact oriented variables are nullified, ease of avoidance of the contact is statistically related to ABS-MI levels 3 (moral evalua- tion), 4 (hypothetical behavior), and 5 (personal feelings) for the schiZOphrenic and levels 2 and 3 for the normal subject groups. It is not related to any of the other 118 I levels for the latter groups or to any of the attitudinal levels for the geriatric and alcoholic subjects. Partial correlations between amount (frequency) of contact and the ABS-MI levels indicate a significant negative relationship on level 2 for the schiZOphrenic group and on levels 1 and 2 for the group of alcoholics. The only positive partial correlations between the criterion and frequency of contact occur on level 6 for the geriatric and normal samples. All partial correlations between the various six attitudinal levels and alternative rewarding Opportunities lacked sta- tistical Significance except for level 6 of the geriatric group. The partial correlation for this lone exception is significant and positive. Thus, the results presented in Table 12 indicate that when the four contact variables are considered individually, reported enjoyment of contact with mentally ill persons generally accounts for more of the variance contained within the statistically significant multiple correlation coeffi- cients than does any other single contact variable for the schizophrenic (except levels 2 and 6), geriatric (except level 6), and alcoholic (except level 1) samples. Data for the normal subject group indicate that, except for personal action level 6, the contact variable designated as ease of avoidance, most often accounts for more of the variance found in significant multiple correlations (levels 3-5) than does any other single contact variable. 119 Examination of the multiple correlations in Table 12 for the total sample reveals small but statistically sig- nificant correlations on levels 1 and 2 as well as compar- atively larger correlations of greater significance on levels 4-6. Although these values do not constitute a test of Hypothesis 9, they do suggest that the combination Of contact variables are indicative of attitudes toward the mentally ill. Straightforward correlational analyses of attitudes toward the mentally ill and the four contact variables taken separately permit further examination of their re— lationship from another perspective. While such correla- tional analyses do not constitute a test of Hypothesis 9, they do provide important results which contribute to the advancement of understanding and knowledge relevant to the relationship between attitudes toward the mentally ill and contact predictor variables. The correlations between all six levels Of the ABS-MI and the amount, alternatives, avoidance, and enjoyment variables are shown in Table 13. One of the most striking aspects of the data is the generally significant positive relationship between reported enjoyment of contact with mentally ill persons and attitudes toward the mentally ill as reflected by scores on the criterion for all groups of psychiatric patients. The relationship is significant, mostly at the .01 level of confidence, for attitudinal 120 .Ho.v.hk« .mo.v.m« oH. oo. oo. «NH. No. «HH.- oHoeem Hence HNN oH. HH. So. NH. mo. oH.- mHesuoz mm oN. HH. SH. oo. mo. oo.- moHHoeooHa om mo>HoeenouHe «SN. So. mo. oo.- mo. SH.I moHnomHnoo Hm oH.- HH. HH. «oN. NH. oo.- moHcmneooNHSom SHH atom. acmm. «awn. ««Hv. ath. So. OHQEdm HMHOB HmN ««NN. ««om. oo. HH. So. NH. mHaSNoz mm «NN. ««oo. ««oo. ««oo. ««NN. oo.- moHHoeooHa om neossoHem «svm. «awe. «com. «com. «adv. No. mOHHHMHHmO Hm oo. ««NN. ««NN. ««NN. ««NN. No. moHsoneooNHaom SHH oo. mo. «NH. ««eH. Ho. No.I mHosam Hence HNN mo. SH. «NN. «HN. oo.- oo. mHmsuoz mm oH. oH. oN. oN. ««Hm. mH. moHHohooHe om outmoHo>< No.I mo.I oH. oH.- oo.- oH. mOHunaHooo Hm oo. NH. «oH. «oH. HH. mo. moHemueooNHeom SHH oH. «NN. oo. mo. oH. oo. oHoesm Hence HoN ««He. SH. oH. oH. oN. So. mHmeuoz mm «NN. no. No. oH.- «oN.- ««No.- moHHotooH4 om ucsoea ««He. NN. NN. NH. oo. oH. moHuumHumo Hm Ho. ««NN. NH. «NH. No. SH.I moHcongoouagom SHH o m o N N H mCOHUGHOHHOU HO>OQ OHMUW HZlmm¢ msouo WOHQMflHU> HUM#GOU .nOHmEmm Hmeuoc can OHHuchommm MOM mcoHumHouuou vomucooIHzImmOH OHmom Hzlmmm msouo .wOHmEom HoEHoc can OHHumHnommm How mcoHHMHOHHOO OHQoHHm> SOMOHHHOIHZImmNII.VH mHmOH OHoom HEImms msouo .mOHmEmm HmEHOS coo OHHuoHnommm How mcoHHOHOHHOO OHQMHHS> ONCOHBOSRIHEImmAlcoholics>Schizophrenics>Geriatrics. Hypothesis 12 was tested by a one-way analysis of variance procedure for each of the samples on each of the ABS-MI scale levels using means adjusted for sample size and sex differences. Also, each mean was tested against every other mean by a multiple means test which generates an E test that produces results equivalent to a two-tailed 3 test. For the readers convenience, the sample groups are listed in the hypothesized order in Table 16 where the 128 .mo.v.m« mOHSOHsQONHnOmv mOHsz nOHHonOOHsm mOHmEomo mHmEuozN mOHuuoHHOOm ONHm OHmEom can xmm How pmumsno¢H dkm .oAm .dAz «mm.v mo.vv mo.vo mo.Nv om.Sv vo.ov mm.mv cOHu04 HocOmuom .o cAz .mAz .oH oHoom qumm< OHHHMHnOSmQ How muHSmou .OHQEmm HmEHoc now who» momma OHQHDHSE pom .m.m H .mcmoe ooumofloo HzImmaII.oH mqmae 129 analysis of variance results appear. InSpection of the data reveals that the hypothesized order of the four groups was not completely achieved on any level of the ABS-MI. In the sense that no contradictions to the hypothesized order occurred, it was partially achieved on moral evaluation level 3 and personal feeling level 5. However, the differ- ences between the means for all of the groups were not Sig- nificant on either level 3 or level 5. There were no sta- tistically significant differences between the alcoholics and either the schizOphrenics or geriatrics on level 3. On the other hand, there were no significant differences be- tween the alcoholics, schizOphrenics and geriatrics on per- sonal feeling level 5. The normals scored significantly higher than other groups on both the former and the latter ABS-MI scales. Contradictions to the hypothesized order of groups occurred on stereotypic level 1, normative level 2, hypo- thetical action level 4, and personal action level 6. The fact that the schiZOphrenic group scored significantly higher than either the alcoholics (levels 1, 2, 4, 6) or normals (level 2), or both (level 2) accounts for the occurrence of the contradictions in all cases. Few significant differences resulted between either the normals or schizophrenics on the one hand, or the alcoholics and geriatrics on the other. These results, in addition to the contradictions already delineated, pre- cluded greater support for Hypothesis 12. 130 Consistencies in the data in Table 16 which are sup- portive of Hypothesis 12, in cases where significant dif- ferences occurred, include the following: (a) normals scored higher than alcoholics, geriatrics, and, with one exception (level 2), schiZOphrenics, and (b) schizOphrenics scored higher than geriatrics. Although somewhat unrelated to Hypothesis 12, it should be pointed out that no statistically significant differences occurred between the males and females on any ABS-MI scale. This finding further supports the justifi- cations listed in Chapter IV for combining the two groups. It may be recalled from Chapter IV that a different pro- cedure (Guilford, 1956, p. 224) failed to reveal any sig- nificant differences between the two sexes on any of the ABS-MI scales or on a number of other variables. Returning to Hypothesis 12, it can be concluded that the hypothesis received the greatest support on ABS-MI levels 3 and 5, but that it was not confirmed in total for any scale. Relating Attitudes and Multidimensionality H-13.--The ABS-MI scale levels or attitude suhf universes will form a Guttman Simplex for each of the sample groupg. Hypothesis 13 was tested by subjecting the combined content-intensity scale level intercorrelation matrices to 131 Kaiser's (1962) simplex approximation test. It will be recalled that Kaiser's test produces a goodness of fit value (Q2) for the obtained matrices and also rearranges these matrices into a "best" simplex order for which a g2 value is assigned. The obtained and empirically reordered matrices, together with their corresponding 92 values, are included in Table 17 for each of various groups. The matrices in Table 17 contain from zero to four statistically non-significant negative correlations. Three occur in the matrix for the schizophrenics, four in the geriatrics matrix, none in the alcoholics matrix, one in the normals matrix and two in the matrix for the total sample. Most of the negative correlations occur in the first two columns Of the matrices and involve correlations between either ABS-MI scale level 1 or 2 and the other scale levels. It was pointed out in Chapter IV that presently, no one has devised a wholly satisfactory solution for the problem created by negative correlations. Therefore, it was decided to treat negative correlations as positive values for purposes of computer computation of g? values and to take this into account in evaluating the results. The figures in Table 17 consist of Q2 values with negative correlations treated as positive correlations. As noted in Chapter IV, Hamersma (1969) alledged that a Q2 value of .70 or greater provided acceptable TABLE l7.--Q2's for Obtained and empirically ordered matrices on for psychiatric and normal sample. six level ABS-MI 1 2 3 4 S 6 1 "‘ 2 2 .29 --- g = .900 Obtained 3 .25 .59 --- Matrices 4 .12 .44 .71 --- 5 .24 .44 .54 .45 --- 1 6 -.15 -.06 .06 .21 -.09 --- 117 Schizophrenics 1 "‘ 2 2 .29 --- Q — .916 Ordered 3 .25 .59 --- Matrices 4 .24 .44 .54 --- S .12 .44 .71 .45 --- 6 -.15 —.O6 .06 -.09 .21 --- 1 "‘ 2 2 .10 --- Q = .845 Obtained 3 -.06 .61 --- ‘ Matrices 4 -.12 .59 .88 --- 5 -.14 .50 .74 .76 --- 2 6 -.12 .23 .52 .54 .58 --- 51 Geriatrics 1 "’ 2 2 .10 --- g = .845 Ordered 3 -.06 .61 --- Matrices 4 -.12 .59 .88 --- 5 -.14 .50 .74 .76 --- 6 -.12 .23 .52 .54 .58 --- l --- 2 2. .36 --- Q = .960 Obtained 3 .26 .56 --- Matrices 4 .26 .56 .87 --- S .33 .54 .72 .79 --- 3 6 .28 .51 .59 .67 .65 --- 58 Alcoholics l --- 2 2 .36 --- Q = .965 Ordered 3 .33 .54 --- Matrices 4 .26 .56 .72 --- S .26 .56 .79 .87 --- 6 .28 .51 .65 .59 .67 --- 1 "' 2 2 .19 --- Q = .857 Obtained 3 -.11 .23 --- ‘ Matrices 4 .02 .21 .87 --- S .21 .29 .60 .62 --- 4 6 .17 .13 .29 .29 .27 --- 55 Normals 1 "' 2 2 .19 --- Q = .884 Ordered 3 .21 .29 --- Matrices 4 -.11 .23 .60 --- S .02 .21 .62 .87 --- 6 .17 .13 .27 .29 .29 --- l "' 2 2 .35 --- Q s .856 Obtained 3 .18 .45 --- Matrices 4 .05 .35 .80 --- 5 .25 .45 .63 .61 --- 5 6 -.17 -.02 .28 .40 .18 --- 281 Total Sample 1 --' 2 2 .35 --- g = .884 Ordered 3 .25 .45 --- Matrices 4 .18 .45 .63 --- s ‘.05 .35 .61 .80 --- 6 -.17 -.02 .18 .28 .40 --- 1Critical value of r at .05 level=.18. 4Critical value of r at .05 1evel=.26. 2Critical value of r at .05 level-.27. 5Critical value of r at .05 1evel=.12. 3Criticalvalue of r at .05 leve1=.25. 133 evidence that a simplex has been approximated. On the basis of that criterion and inasmuch as all of the g2 values in Table 17 are .845 or higher, it is concluded that a simplex has been approximated for all of the various groups of subjects. These results are in direct support of Hypoth- esis 13. The greatest stock can be placed in the simplex for the alcoholic group because: (a) 92 values of .960 and .965 for the obtained and ordered correlation matrices respectively, exceed the .70 criterion by far and are the highest for this group; (b) all of the correlations within the matrix are statistically significant at the .05 level; and (c) all correlations within the simplex are positive values. The simplexes for the geriatric group may be seen as earning the least credence and greatest number of quali- fications. The Q? values of .845 for both the obtained and ordered matrices are the lowest for any group and, in addition, four negative correlations resulted. The total effect of the negative values for this group is unclear, but it may be tempered somewhat by the fact that none of them are statistically significant, all appear in column one, and the column generally exhibits descending correla- tion coefficients moving from tOp to bottom. The results in Table 17 provide supportive evidence for Guttman's (1959) Contiguity Hypothesis which states 134 that attitude levels closer to each other in terms of the semantic scale of their definitions will also be closer statistically. In other words, the resulting correlation matrices reveal what Guttman (1966) has termed a "simplex" ordering and the 92 values in Table 17 indicate that a simplex has been approximated for each group of subjects within this research. Finally, the relatively small dif- ferences between the 92 values for the ordered and obtained matrices for each group suggests a close correspondence between the hypothesized order and the "best" order or between the semantic meaning specified in the six levels on the obtained statistical structure. CHAPTER.VI SUMMARY, DISCUSSION, AND RECOMMENDATIONS Attitudes related to the mentally ill have been con- sidered primarily in terms of how parents, lay public, and professional mental health caretakers view and/or influence those who are mentally ill. Very little systematic atten- tion has been given to how psychiatric patients View men- tally ill persons or what impact they might have on one another. Problem The reality of the truly interdependent nature of attitude formation and eXpression on the part of one psy- chiatric patient for another has been all but overlooked. Such a reality continues to exist despite recognition in other areas of the part played by intergroup attitudes in predicting, understanding, and controlling human relations. It continues also in spite of the importance of psychiatric patient attitudes insofar as they relate to social struc- tures, conceptualizations of mental illness, develOpment of treatment programs, and the develOpment of training and educational programs for mental health workers. 135 136- In addition, it is not generally known what impor- tance can be attributed for different attitudinal levels to: (a) the amount of contact a person has had with the mentally ill, (b) the value orientation of the person, (c) the amount of factual knowledge about mental illness he possesses, and (d) the demographic characteristics of the subjeCt. Consequently, the attitudes of the mentally ill on a comparative basis are far from clear and are in need of research designed to delineate their structure, content, determinants, and relationships. Thus, the major purposes of the present research were the following: 1. To construct an attitude scale (ABS-MI) accord- ing to the formulations of Guttman (1957, 1959, 1966), Jordan (1968), and Maierle (1969), with the mentally ill as the attitude object. 2. To measure and determine predominate value orientations and attitudes toward the mentally ill of normal and psychiatric persons. 3. To determine the predictive ability of hypoth- esized determinants (demographic, valuational, contactual, and knowledge) of attitudes toward the mentally ill. 4. To compare a group of normals and samples of hospitalized psychiatric patients )schizophrenics, geri- atrics, and alcoholics) on the basis of their attitudes 'toward the mentally ill, several predictor variables, and 137 the interrelationship between attitudes and independent variables. 5. To test Guttman's Contiguity Hypothesis which specifies that responses to an attitude scale develOped on the basis of facet analysis will form a matrix of level- by-level correlations which will approximate a simplex (Guttman, 1959, 1966). 6. To examine the effects of order of attitude scale levels on simplex approximations and the correspondence between a theoretical order of administration of scale levels and a "best" order for psychiatric and normal persons. Related Research Research concerned with attitudes of schizophrenics, geriatrics, alcoholics, and normals toward the mentally ill was reviewed. Areas reviewed included attitudes of self, peers, community, students, young persons, lay groups and caretakers. A review of the literature revealed that a great deal of discrepancy existed regarding how psychiatric patients and normals perceive persons who are mentally ill. It was reported in some studies that the image of the mental pa- tient as seen by both patients and normals is rather un- flattering. However, other studies related that psychiatric patients' attitudes toward the mentally ill were no more devient than were those of Staff members in the fields of 138 psychiatry, psychology and psychiatric social work. Still other research suggested that attitudes held by normal subjects were not really "bad." Research literature concerning attitudes toward the alcoholic is limited in both number and sc0pe. What exists indicates that alcoholism, like mental illness, has been negatively viewed by the general public, students, and caretaker groups such as physicians and social workers. Few attemtps have been made to evaluate attitudes of aged patients toward the mentally ill. Instead, attitudes of elderly individuals, young people, and psychiatric pa- tients toward the aged and aging have been examined in a variety of studies. The results indicate that all three groups tend to have a rather negative appraisal of older peOple and generally view old age in a gloomy manner. Although the literature indicated that numerous variables concerning contact, demographic, knowledge, and value orientation factors were related to attitudes, most of the research was inconclusive or contradictory with reference to predictor variables. The reason for this phenomenon might be attributed to the fact that various measures of attitudes were composed of items derived from different attitudinal structures and, therefore, contri- buted to a lack of control over the various attitudinal levels being measured. None of the studies used an atti- tude scale based on the structural facet theory proposed by Guttman. 139 The review of the literature in Chapter II also in- cluded a review of facet theory and methodology. A synop- sis of that review follows. Guttman identified four levels, or types of attitudes, generated from permutations of three facets which struc- ture the attitude-object relationship. He formulated a Contiguity Hypothesis which specified that the matrix of attitude level-by-level correlations will approximate a simplex. Jordan extended Guttman's analysis to a five facet, six-level system and found that data derived from a scale on attitudes toward the mentally retarded (ABS-MR) did generate simplex approximations. Maierle analyzed 32 permutations of five dichotomized facets and found that only 12 appeared semantically pose sible. He hypothesized that these 12 members were ordered within seven semantic paths, or ordered sets of level— member items. A set of experimental instruments was develOped and subsequently administered to a large group of subjects in a random order and in a hypothesized order. Maierle's results indicated the lack of an ordering prin- ciple obviously better than the one hypothesized by Jordan and a generally close correspondence between hypothesized and best orders. 140 Instrumentation This study employed a new criterion instrument--the Attitude Behavior Scale-Mental Illness (ABS-MI) which was evolved by a Guttman facet theory and analysis rationale. It contains six levels of attitude structure: (a) Societal Stereotype; (b) Societal Norm; (e) Personal Moral Evalua- tion; (d) PersOnal Hypothetical Action;(e) Personal Feelings; and (f) Personal Action. Eight content items, each with a corresponding mea- sure of intensity were selected for each of the six levels of the instrument so that the final attitude scale con- sisted of 96 items. Thus, each one of eight attitude items and a corresponding measure of intensity was repeated across all six attitude levels with the attitude items being altered to conform to the facet structure on the different levels. The eight content areas are: (a) Marriage-to; (b) Intelligence-of; (c) Understanding-of; (d) Invitations-to; (e) Friends-with; (f) Eating-with; (g) Helpfulness-of; and (h) Loans-to. The content and intensity measures were combined for each item on the ABS-MI except when the relationship be- tween attitudes toward the mentally ill and frequency of contact was analyzed. The content and intensity values for each item were combined also on the Efficacy Scale. The rationale and procedure used for combining the two vari- ables was provided in Chapter III. 141 Various aspects concerning the validity of attitude behavior scale type instruments were discussed in Chapter IV and it was emphasized that they provided a rather novel and unique opportunity to compare different dimensions of atti- tudes. It was pointed out that the use of facet theory and a mapping sentence assures that adequate content valid- ity of attitudinal levels is maintained. In addition, a variety of researchers, including the present one, have obtained simplexes for a wide range of subjects as pre- dicted by Guttman's facet theory. These results provide evidence of construct validity for attitude behavior scales that have been constructed on the basis of facet theory. Finally, attitude behavior scales have been validated through the use of contrasted groups and researchers have demonstrated considerable evidence of "correlational" validity in that a number of predictor variables and mea- sures of attitude have tended to evolve as hypothesized. Kuder-Richardson-like reliabilities for three samples of psychiatric persons and a sample of normals on six levels of the ABS-MI ranged from .54 to .86. As a whole, the reliabilities were comparable to those obtained on other attitude scales. ‘ Attitude content and intensity scores were combined into one for each item as previously indicated. The result was to increase the majority of reliability estimates across groups and ABS-MI levels. 142 Measures of independent variables were contained within a 54 item questionnaire. Among the items were mea- sures of (a) demographic variables, (b) change orientation, (c) chronicity orientation, (d) contact with the mentally ill, (e) efficacy, or man's effectiveness in the face of his natural environment, and (f) knowledge about mental illness. Design and Analysis The ABS-MI and an accompanying questionnaire was administered to three samples of hospitalized psychiatric patients as well as to a group of normal persons. The samples consisted of (a) 117 schiZOphrenics, (b) 51 geri- atrics, (c) 58 alcoholics, (d) 55 normals. The psychiatric patients were all hospitalized and housed on Open, unlocked wards throughout the Battle Creek, Michigan, Veterans Administration Hospital. The sample of normals consisted of graduate students in rehabilitation counseling or per- sonnel and guidance services at Michigan State University. A number of procedures were used to test the various hypotheses. These included product—moment, partial, and multiple correlation procedures as well as one-way analysis of variance and a multiple means test. Also, a simplex approximation test was used which produces a descriptive statistic (g2) fer obtained attitude level matrices and matrices reordered into a "best" simplex order. 143 A total of 13 research hypotheses were formulated which were based on previous research. A summary of the results of the testing of each hypothesis will be pre- sented in the following segment. Results Reliability estimates were obtained for each of the sample groups on each of the ABS-MI levels for content scores as well as combined content-intensity scores to test the power of this procedure in the present research. The reliabilities for the combined scores ranged from .55 to .86 and the majority were consistently higher than the content-only reliabilities which ranged from .54 to .81. A summary of the results of testing each hypothesis follows. H-l.--Persons who score high on age will score low on positive attitudes toward the mentally ill. Hypothesis 1 predicts a negative relationship be- tween age and attitudes toward the mentally ill. The hypothesis was partially supported in that high age and positive attitudes formed a negative relationship on levels 1 and 3-5 for the total sample, levels 3-4 for both schiZOphrenics and normals; and level 1 for the geri- atrics. The data for the alcoholic subjects did not sup- port Hypothesis 1 and in fact, the results obtained for levels 1 and 2 for this group were contrary to those pre- dicted. 144 H-2.--Amount of education will be positively related to favorable attitudes toward the mentally ill. Amount of education was found to be positively re- lated to ABS-MI scores on levels 3-6 but negatively related on normative level 2 for the total sample. These apparently conflicting results were interpreted as indicating that increasing education was related to a tendency to attri- bute relatively poor attitudes to others on a normative level and positive attitude to oneself on levels 3-6. Of all the individual groups, Hypothesis 2 was sup- ported most by the data obtained on levels 3-4 and 6 for the group of normals. H:§,--Persons who have experienced a relatively short period of hospitalization will score higher on positive attitudes toward the mentally ill than will persons who have experienced a relatively long_period of hoSpitalization. Shortness of hospitalization was positively related to attitudes toward the mentally ill on ABS-MI levels 2-3 and 6 but a negative relationship occurred on normative level 2 for the total sample. Thus, persons whose hos- pitalization had been of shorter duration attribute more favorable attitudes to themselves and less positive atti- tudes to others. Considering the groups separately, Hypothesis 3 was confirmed on moral evaluation level 3 and hypothetical action level 4 for the schizophrenics and on levels 3-5 145 for the normals. Results obtained for the alcoholic and geriatric group failed to support the hypothesis. It was pointed out that the data for the total sample and normals should be interpreted with caution since the normals may have adopted an inappropriate frame of reference while responding to the particular item con- cerning length of hospitalization. H:4.--Persons who score high on stated importance of religion will score low ongpositive attitudes toward_ the mentally ill. A negative relationship between stated importance of religion and positive attitudes, as predicted by Hypoth- esis 4, was not found on any ABS-MI levels for any group of subjects. Therefore, it was concluded that the results demonstrated a lack of support for the hypothesis. H:§.--Persons who score high on stated adherence to religion will score low on positive attitudes toward the mentallypill. A statistically significant negative relationship between the criterion and adherence to religion occurred on stereotypic level 1 for both the total sample and the schiZOphrenics. However, a positive relationship resulted on levels 3-4 and 6 for the schizophrenics whereas the hypothesis predicted a negative one. It was concluded that Hypothesis 5 was unsupported. 146 H-6.--Persons who score high on change orientation will score high on positive attitudes toward the mentally ill. The relationship predicted by Hypothesis 6 was largely supported for the total sample as well as for the geriatric, normal, and alcoholic groups. It was found that individual change orientation variables appeared to relate differentially to attitudes and to vary according to subject group and ABS-MI scales. The best sample group-predictor variable combinations were: (a) geriatrics-self change, (b) normals-birth control, and (c) alcoholics-child rearing. H-7.--Persons who score high on chronicity orienta- tion will score high on positive attitudes toward the mentally ill. High scores on the chronicity variable indicate rejection of an orientation toward chronicity. Therefore, a positive relationship between the two variables suggests that positive attitudes toward the mentally ill are asso- ciated with lpw_chronicity orientation. Hypothesis 7 is supported to a great extent for the total sample as well as for the schiZOphrenics, geriatrics, and normals. It is supported only on the more impersonal normative level 2 for the alcoholic sample. The strongest single predictor of attitudes was work involvement for the total sample, schiZOphrenics, and 147 geriatrics, whereas responsibility was most powerful for the normal group. A single "best" predictor variable failed to emerge for the alcoholics. H:§.--The more frequent the contact with mentally ill persons the higher will be the intensipy scores on the ABS-MI, regardless of the direction lpositivepess or nggativeness) of attitude. Frequency of contact was found to be related to atti- tudes toward the mentally ill for the total sample. Hypoth- esis 8 is also supported for the alcoholic sample but it was not confirmed for any of the other three individual groups. H:2.--High frequengy_of contact with mentally ill persons will be associated with favorable attitudes toward the mentally ill if high fregpengy is concurrent with (a) alternative rewarding Opportunitiesy (b) ease of avoidance of the contact, and (c) epjpyment of the contact. Hypothesis 9 could not be tested directly as origi- nally planned since the data did not meet Specified con- currency conditions. However, multiple, partial, and product-moment correlations were obtained which revealed several significant relationships between contact variables and attitudes toward the mentally ill. The most salient of these were the following: (a) contact variables con- cerned with avoidance, amount, alternatives, and enjoyment, ‘when combined, were significantly related to attitudes 148 toward the mentally ill for all subject samples; (b) enjoy- ment of contact with mentally ill persons was the strongest predictor of attitudes; and (c) alternative rewarding opportunities was the weakest predictor of attitudes. H;l9.--Persons who score high in efficacy will score high in positive attitudes toward the mentally ill. By and large, Hypothesis 10 was confirmed for the schiZOphrenics but it was not supported for the geriatrics and alcoholics. It was supported on moral evaluation level 3 and hypothetical action level 4 whereas it was unsupported and contrary to prediction on stereotypic level 1 for both the normals and total sample. H:ll.--Persons who score high in knowledge about mental illness will score high in_positive attitudes toward the mentally ill. Knowledge about mental illness was significantly related to attitudes toward the mentally ill on ABS-MI levels 2-6 for the total sample and on levels 2-5 for the schizophrenics, providing direct support for H-ll. The relationship emerged on only personal action level 6 for both the geriatrics and alcoholics. H:lll--The gropps will assume the following order with reSpect to favorable attitudes toward the mentally ill: Normals>A1coholics>SchiZOphrenics>Geriatrics. Hypothesis 12 received the strongest support on ABS-MI moral evaluation level 3 and personal feeling level 5; 149 however, it was not totally confirmed on any attitudinal level. Even on levels 3 and 5, not all of the means dif- fered significantly from one another. Consistencies in the data that were supportive of H-12 were: (a) normals scored significantly higher than alcoholics, geriatrics and, with one exception on level 2, schizophrenics, and (b) schizophrenics scored higher than geriatrics. H-13.--The ABS-MI scale levels or attitude sub-' universes will form a Guttman Simplex for each of the samplelgroups. Using a 92 criterion value of .70 or greater, a simplex was approximated for normal, alcoholic, schizo- phrenic, and geriatric samples as predicted by Hypothesis 13. Also, a close cOrrespondence resulted between the hypoth- esized order and "best" order. Discussion of Results A review, interpretation, and integration of the major findings of the present research is included in this section. ABS-MI Reliability Reliabilities for the six levels of the ABS-MI were generally satisfactory and were comparable to other atti- tude scales. Consistent increments in the majority of the reliabilities of the scale levels occurred as a result of combining content and intensity scores. The effectiveness 150 of the procedure was greatest for the lowest reliabilities which generally showed the greatest overall increase from combining content-intensity scores. The pattern of the reliability coefficients vary for the three psychiatric samples, but in general they show a distinct difference when compared with those obtained for the normal group. The reliabilities tended to be highest on the more impersonal ABS-MI levels and lowest on the more personal feeling level 5 and behavioral action level 6 for the psychiatric groups. The highest reliabilities for the normal subjects were obtained on scale levels 4-6, while the lowest resulted on levels l-2. The most adequate interpretation of these results is that the psychiatric subjects were more certain and con- sistent with regard to how other people compare the men- tally ill and how other people behave toward this group than they were of how they themselves personally feel and act toward the mentally ill. On the other hand, the normals were less consistent or sure about how other peOple should act towards and compare persons who are mentally ill. Normals were more consistent and positive with re- spect to how they would act toward mentally ill persons and how they personally felt and behaved toward them. Relating Attitudes and Demographic Variables Age was related to ABS-MI attitudes in that older persons in the tOtal sample expressed more negative attitudes 151 toward the mentally ill on an abstract and impersonal stereotypic level as well as on progressively more concrete and personal levels ranging from personal moral considera- tions to personal feelings. As a whole, the more aged subjects tended to believe that other people maintained negative attitudes about the mentally ill. Also, older persons were of the opinion that they should not and would not interact positively with mentally ill persons, and that they had rather negative personal feelings about them. The relationship between age and attitudes varied within and between groups. Schizophrenics and normals who were older emitted more negative attitudes on both a personal moral and a hypothetical behavioral level, whereas older geriatrics attributed more unfavorable attitudes to others. Otherwise, the older members of the geriatric group did not express unfavorable attitudes toward the mentally ill any more than did those of their group who were younger. The results for the alcoholics were somewhat con- trary to those of the other groups since older alcoholics ascribed favorable attitudes to others on an abstract, im- personal attitudinal level. Thus, increasing age is related to attitudes toward the mentally ill but the relationship is one that varies according to subject groups and attitudinal structure. 152 Amount of education was found to be positively related to attitudes toward the mentally ill on the increasingly more personal and action oriented attitudinal dimensions for the total sample. On a less personal, societal norma- tive, amount of education and attitudes were negatively related. The two variables, age and education, were un- related for the geriatrics and alcoholics and unrelated on a stereotypic level only for the schizophrenics. They were related for the alcoholics on the moral evaluation, hypothetical and personal action dimensions of the ABS-MI. The occurrence of a significant negative relationship on the normative level and significant positive relation- ship on the remaining attitudinal levels is an event re- quiring special mention and discussion. This event, or one in which the relationship between attitudes and a par- ticular variable turns out to be significantly negative on either levels 1 and 2 (stereotypic and normative), or both, while positive relationships result between the variables On one or any number of the remaining levels (3-6), or vice versa, occurred quite frequently throughout the results of this research. As noted by Harrelson (1970), these opposite-direction correlations can be attributed to the fact that the Referent (Facet A), it will be recalled from Figures 3-6 in Chapter III, shifts from Others (al) to Self (a2) from levels 2 to 3 in the ABS-MI. The effect of this shift is that subjects frequently attribute more 153 negative attitudes to others and more positive attitudes to themselves and in some instances, just the reverse. Speci- fically in the case of the relationship between education and attitudes, better educated subjects in the total sample attributed less favorable attitudes toward the mentally ill to others, whereas they identified themselves as having more positive attitudes. A relatively short duration of hospitalization was related to positive attitudes on a personal moral evalua- tion and affective feelings level for the schizophrenics. Attitudes and length of hospitalization were correlated for both the geriatric and alcoholic sample. Most of the alcoholics had been hospitalized for only a short time. Therefore, this variable is relatively homogeneous within this group which may partially account for the lack of any significant relationship between attitudes and length of hOSpitalization in their case. Although attitudes and length of hospitalization were positively related on a number of levels for both the nor- mals and the total sample, it was pointed out that these results may be falacious and should be interpreted cautiously since the normal group may or may not have considered only their non-psychiatric hospital experiences when responding to the item concerning this variable. 154 Relating Attitudes and Religiosipy Variables concerning religious adherence and religious importance were not strong predictors of attitudes in either the psychiatric or normal samples. Generally the predicted negative relationship between these variables and ABS-MI attitudes failed to materialize. Earlier, a predicted negative relationship between the religious variables and attitudes toward the mentally retarded (Harrelson, 1970) and attitudes toward the physically handicapped (Jordan, 1968) failed to emerge. In view of the results of these studies as well as those of the present one, religious im- portance and adherence cannot be considered as strong cor- relates or determinants of attitudes toward the mentally ill, mentally retarded, or physically handicapped. Relating Attitudes apd Change Orientation The relationship between change orientation and atti- tudes was inconsistent in that the relationship varied for subject groups, ABS-MI levels, and particular change orientation items. The two variables were related for the groups as a whole; however, an inconsistent mixture of ;positive and negative partial correlations on stereotypic level 2 presented a consistent interpretation of the data for this level. Interestingly, individual change orientation items related differentially to attitudes within samples and 155 between groups. For example, a self-report measure of "ease of self change" was the strongest predictor of atti- tudes for the geriatrics, but it was not a good predictor for the remaining groups. On the other hand, a measure of beliefs regarding birth control was the strongest predictor of attitudes toward the mentally ill for the normal sample. In general, the data indicated that a combination of change orientation items was predictive of attitudes and that some individual items were better predictors than others for some groups; however, no single item emerged as the strong- est predictor for all samples. Thus, the multiple correlation procedure, which was employed to examine the relationship between attitudes and change orientation predictor variables, proved to be a valuable tool in that it helped to identify relationships which might have otherwise gone unnoticed. Relating Attitudes and Chronicity Orientatigp The pattern of the relationship between chronicity orientation variables and attitudes toward the mentally ill is similar to that described in the previous section regard- ing attitudes and change orientation. That is, individually and in combination, chronicity orientation variables are predictive of attitudes; however, the relationships vary in terms of strength depending upon particular groups, ABS-MI levels, and predictor variables. 156 Although no single predictor was "best" for all samples across all ABS-MI levels, the item which pertained to a willingness to become involved in activities related to work was predictive of attitudes for more groups across more levels than any other variable. The value or impor- tance of the two variables concerning opinions regarding self responsibility for one's own behavior and the need to earn privileges is at best limited, at least for the psy- chiatric samples. In future research of this type, the elimination of these variables might be considered. Relating Attitudes and Contagt Frequency of contact with mentally ill persons was unrelated to certainty or sureness of attitudes toward them for any sample except the alcoholics. The more contact these subjects had with mentally ill persons, the more definite they were with respect to their attitudes. Furthermore, increased amount of contact was associated with a tendency for the alcoholic to believe that the atti- tudes of others toward the mentally ill were quite nega- tive whereas his own were positive with respect to inter- acting with them on an actual behavioral level. It would seem that the more an individual reported he enjoyed his contacts with mentally ill persons, the more positive would be his attitudes toward them. Such a rela- tionship is precisely the case for all groups of psychiatric patients. Whether schiZOphrenic, alcoholic, and geriatric 157 persons enjoy their contact with the mentally ill is, by far, the strongest predictor of their attitudes toward them. The schizophrenics indicated that their attitudes toward the mentally ill were more favorable if they had had some opportunity to avoid such contacts. The same rela- tionship, only to a lesser extent, occurred for the normal group. It may be that the former feel satiated with ex- cessive contact, while the latter have experienced few contacts and are fearful of them. Relating Attitudes and Efficacy Efficacy was a strong predictor of attitudes toward the mentally ill for the schizOphrenics. Those who per- ceived man as able to COpe effectively with various aspects of his natural environment, expressed more positive atti- tudes. On a moral evaluation, and a hypothetical behavior level, efficacy was predictive of attitudes for the normal sample. For this group, however, as well as for the geri- atrics and alcoholics, a sharp differentiation on the ABS-MI occurs between levels l-2 and 3-6 where the "referent" shifts from others to "self." On the first, second, or both levels, subjects who scored high on the efficacy scale assigned less favorable attitudes to others. Generally, the efficacy variable was not a strong predictor of attitudes toward the mentally ill on the more personally oriented action levels of the criterion, es- pecially for the geriatrics, alcoholics, and normals. 158 Relating Attitudes and Knowledge Knowledge about mental illness is a good predictor of attitudes toward the mentally ill for schizophrenics and normals. Also, if a slightly less than .05 level of sig- nificance was selected, it would be a reasonably strong predictor for the alcoholics on ABS-levels 3-5 as it already is on the more personal and concrete actual behave ioral level 6. With reSpect to the geriatric sample, knowledge is a strong predictor of the occurrence of favorable personal interactions with the mentally ill. Generally the knowledge variable is unrelated to the more abstract and impersonal ABS-MI levels for the geriatrics and alcoholics. In effect, the more knowledgeable geriatrics failed to indicate that they ought to interact with the mentally ill or that they felt very positive toward them; however, what they did indicate was that they have interacted with them on a number of personal levels. On the other hand, the normals and the schizophrenics suggest that they should and would interact with the mentally ill and have positive feelings about them, but they either do not or have not interacted together. Initially it was somewhat confusing and puzzling to find that schizophrenics who are more knowledgeable about mental illness, say that they experienced little inter- personal involvement with other persons who were mentally 159 ill. One explanation of this phenomenon might be that although schizophrenics are in the physical presence of mentally ill persons almost continuously, they do not tend to be within one another's psychological or inter- personal proximity or influence. In their daily physical associations together, many of them make it a practice to reject or avoid establishment of personal and emotional contacts and seldom even acknowledge each other in any manner. RelatinglAttitudes and Group Membershlp It will be recalled that the hypothesis relating attitudes and group membership did not receive complete support on any ABS-MI level and that a number of contra- dictions to the hypothesis occurred. When the hypothesis was originally formulated, intui- tively it was believed that responses to the ABS-MI and the items in the questionnaire by alcoholics would be most similar to normals. That belief turned out to be erroneous. In fact, in most instances involving a relationship be- tween a predictor variable and attitudes toward the men- tally ill, the pattern presented by alcoholics was very unlike that of normals. In terms of adjusted mean scores on the ABS-MI (Table 20 in Appendix B), those obtained by the schiZOphrenics are more similar to the adjusted mean scores of normals than are those of alcoholics. This 160 result was significantly contrary to what was predicted on ABS-MI levels 1, 2, 4, and 6. In no case were the mean scores for the alcoholics significantly higher than those of schizophrenics. One further observation that warrants elaboration is that the differences between the means for various groups on the 6 levels of the ABS-MI, often are no greater, or even less, than the number of instances where no signifi- cant differences occur. The findings that schizophrenics tend to express more favorable attitudes toward the mentally ill than do alcoholics, and that no significant differences between the mean attitude scores occur for some groups on various attitudinal levels, has at least one important implication for future research of this nature. With respect to the mean scores on the ABS—MI scales, the hypothesized order of the groups should be the following: normalsischizo- phrenicsialcoholicsigeriatrics. Relating Attitudes and Multidimensionality As predicted by Guttman's (1959) Contiguity Hypoth- esis, a simplex was approximated for the sample of normals and for each sample of psychiatric patients. This finding is supportive of the proposition that there exists an in- variate structure of attitudes toward the mentally ill across different groups of psychiatric patients and normals. 161 Maierle (1969) commented that further research may be needed to establish that the discriminating ability necessary to distinguish among varying level members is found in the less SOphisticated. The present research establishes that the less sophisticated, as well as the more SOphisticated, do possess the necessary discriminatory ability. Recommendations for Further Research A number of recommendations concerning sampling, design, and analysis suggested by Harrelson (1970) and Maierle (1969) were incorporated in the present research. However, various limitations and shortcomings became evi- dent as the study progressed toward completion. A brief summary of their implications as well as recommendations for future research are presented in the final section. Criterion Instrumentation Many psychiatric patients commented about the length of the ABS-MI and Personal Questionnaire. Although most of the schizophrenic and alcoholic subjects cooperated quite well with the administration and completion of the instruments, some of them offered observations and comments indicative of an initial reaction of dissatisfaction re- garding its total length. Many geriatric subjects thought the instrument too long and became impatient, restless, and distracted during the course of testing, and consequently, 162 rapport between them and the examiner was difficult to establish, maintain or both, as was motivation to complete the instrument. It will be recalled that each content item on the ABS-MI Efficacy scale is followed by an intensity item that asks the subject to indicate how sure he is of his response. For many subjects, selection of an intensity item became a rather routine procedure in that they tended most often to be either "fairly sure" or "very sure" of their response to a content item. The experiences and observations mentioned above resulted in the following recommendations: 1. The total length of both the ABS-MI and the Personal Questionnaire should be reduced in future research. 2. The ABS-MI and Efficacy scale should be con- structed and designed so that, if the subject desired, the necessity of indicating an intensity preference fol- lowing eagh content item would be eliminated. A possible alternative might be to ask the subject, at the end of each ABS-MI scale level, how sure he was of his answer. Independent Variable Instrumentation At the conclusion of his research, Harrelson (1970) reported that a need exists for more adequate predictor variables and suggested that perhaps such variables existed in the form of available psychological tests which could be 163 incorporated into future research studies. The same can be said of many of the independent variables used in the present research, even though some of Harrelson's recom- mendations were followed such as combining content and intensity scores into one on the Efficacy scale and re- vising certain demographic variable questions. Therefore, it is recommended that certain sections of the Personal Questionnaire be revised further. The following suggestions are offered for future research. 1. Select an improved measure of religious orienta— tion or eliminate the variable altogether since it was not at all predictive of attitudes in the present research or in that completed by Harrelson (1970) and Jordan (1968). 2. Revise the chronicity orientation items as they are presently constructed. They encompass elements of high social desirability and fakeability and may be in- valid. 3. With regard to the contact variables, correlate the enjoyment of contact variable directly with attitudes. Revising or discarding the avoidance and alternative variable should be given consideration. 4. Select relevant independent variables for which cflojective physiological or observational measures are anvailable or can be obtained. This suggestion applies particularly to change orientation, contact, and chroni- city orientation variables. 164 Analysis The results of the present study reinforce the fol- lowing two recommendations made by Harrelson (1970). The recommendations which appear in a slightly revised form are: 1. Future studies using ABS type instruments as a criterion should employ a two-way analysis of variance procedure to analyze group-scale level interactions. 2. A test of significance is needed for a simplex approximation statistic. The following additional recommendation that emerged from the present research can be added to those listed above. 3. Analysis that provides data for a total group consisting of two or more distinct sub-groups should be carefully evaluated and possibly omitted in future re- search since, in the present study, various total means, scores, and correlational values for combined groups could consist of strikingly different values for the sub-groups. Hypotheses The present construction and form of Hypothesis 9 which relates contact to attitudes and Hypothesis 12 which relates attitudes and group membership does much to pre- clude obtaining the significant relationships they specify. Thus, it is recommended that the hypotheses be re- formulated to read as follows for Hypothesis 9 and l2 respectively. 165 1. Persons who score high in enjoyment of contact with the mentally ill will score high in attitudes toward mentally ill persons. 2. The groups will assume the following order with respect to favorable attitudes toward the mentally ill: Normals:SchiZOphrenicsiAlcoholicsiGeriatrics. Administration As indicated earlier, during the planning stages of this study, it had been decided to administer the instru- ment to the subjects in groups. Subsequent experiences, however, led to group administration to normals, alcoholics, and schiZOphrenics and individual administration to geri- atrics. In future research involving geriatric subjects, individual administration of the instrument to all subjects is recommended. Generalizability The results of the present study can be safely gener- alized only to Caucasian male schizophrenics, alcoholic, and geriatric subjects who reside on unlocked wards at the Battle Creek, Michigan, Veterans Administration Hospital and to a particular group of graduate students at Michigan State University, To insure wider generalization of the results, it is recommended that future research in the area of attitudes toward the mentally ill incorporate a greater diversity of groups in a more extensive geographical area. 166 Theory To data, quite a number of studies, in addition to the present one, have demonstrated that facet theory and analysis provide a very functional and fruitful approach to ordering disparate and seemingly unrelated psychologi- cal events and concepts on a theoretical, predictive, and descriptive level. The greatest weakness in the present research is the lack of adequate attitude predictor variables. In conclusion, the following recommendations are presented. 1. The theoretical basis of the relationships be- tween attitudes and attitude predictor variables, as well as empirical measures of the latter, should be strengthened. 2. Facet theory should incorporate the occurrence of negative relationships between various levels of an attitude universe. 3. To facilitate exploration and identification of attitude—behaviors and personality interrelationships, facet theory and analysis should be incorporated into the construction of both attitudinal and personality measures in future descriptive and experimental studies. 167 REFERENCES Abrams, H. S., & McCourt, W. F. Interaction of physicians with emergency ward alcoholic patients. Quarterly Journal of Studies on Alcohol, 1964, 25, 679-688. Anastasi, A. Psychological testing. (2nd ed.) New York: Macmillan, 1961. Bastide, R., & van den Berghe, P. Stereotypes, norms, and interracial behavior in San Paulo, Brazil. American Sociological Review, 1957, 22, 689-694. Blane, H. Attitudes, treatment, and prevention. In J. H. Mendelson (Ed.), Alcoholism. Boston: "Little, Brown and Company, 1966. Blane, H. T., & Hill, M. J. Public health nurses speak up about alcoholism. Nursing Outlook, 1964, 12, 5. Brady, J. P., Zeller, W. W., & Reznikoff, M. Attitudinal factors influencing outcome of treatment of hospitalized psychiatric patients. Journal of Clinical and Experimental Psychopathology, 1959, 20, 326-334. Carp, F. M. Attitudes of old persons toward themselves and toward others. Journal of Gerontology, 1967, 22, 308-312. Caudill, W.; Redlich, F. C.; Gilmore, H. R.; & Brody, E. G. Social structure and interaction processes on a psychiatric ward. American Journal of Orthopsyghiatry, 1952, 22, 314-334. Chafetz, M. E., & Demone, H. W. Alcoholism and society. New York: Oxford University Press, 1962. Clark, A. W., & Binks, N. M. Relation of age and education to attitudes toward mental illness. PsycholOgical Reports, 1966, 19, 649-650. Cohen, J., & Struening, E. L. Opinions about mental ill- ness in the personnel of two large mental hospitals. Journal of Abnormal and Social Psychology, 1962, 64, 349-360. 168 169 Crumpton, E., Weinstein, A. D., Acker, C. W., & Annis, A. P. How patients and normals see the mental patient. Journal of Clinical Psychology, 1967, 23, 46-49. Crumpton, E., & Wine, D. B. Conceptions of normality and mental illness held by normal and schiZOphrenic adults. PsychiatryiDigest, 1965, 26, 39-43. Drake, J. T. Some factors influencing students' attitudes toward older people. Social Forces, 1957, 35, 266-271. Dubois, P. H. An introduction to psychologlcal statistics. New York: Harper & Row, 1966. English, H. B., & English, A. D. A comprehensive dictionary of psychological andlpsychoanalytical terms. New York: DAvid McKay Company, 1958. Erb, D. L. Racial attitudes and empathy: A Guttman facet theory examination of their relationships and deter- minants. Unpublished doctoral dissertation. Michigan State University, 1969. Farina, A., & Ring, K. The influence of perceived mental illness of interpersonal relations. Journal of Abnormal Psychology, 1965, 70, 47-51. Feifel, H. Psychiatric patients look at old age. American Journal of Psychiatry, 1954, 111, 459-465. Festinger, L., & Katz, D. Research methods in the behavioral sciences. New York: Dryden Press, 1953. Foa, U. G. The contiguity principle in the structure of interpersonal relations. Human Relations, 1958, 11, 229-238. Foa, u. c. A facet approach to the prediction of common- alities. Behavioral Science, 1963, 8, 220-226. Freed, E. X. Opinions of psychiatric hospital personnel and college students toward alcoholism, mental illness, and physical disability. Psycholoqical Reports, 1964, 15, 615-618. Giovannoni, J. M., & Ullman, L. P. Conceptions of mental health held by psychiatric patients. Journal of Clinical Psycholoq , 1963, 19, 398-400. 170 Gottlieb, K. A Guttman facet analysis of attitudes toward mental retardation in Colombia: Content, structure and determinants. Unpublished doctoral dissertation. Michigan State University, 1970. Guilford, J. P. Fundamental statistics in psychology and education. New York: McGraw-Hill, 1956. ment. In S. A. Stouffer (Ed.), Measurement and prediction. Princeton: Princeton University, 1950. Guttman, L. The problem of attitude and Opinion measure- FT Guttman, L. An outline of some new methodology for social : 1 research. Public Opinion Quarterly, 1955, 18, 395-404. 1 j Guttman, L. A structural theory for intergroup beliefs ' and action. American Sociological Review, 1959, 24, 318-328. Guttman, L. The structuring of sociological spaces. Technical note No. 3, 1961, Israel Institute of Applied Social Research, Contract No. AF 61 (052) - 121. United States Air Force. Guttman, L. Order analysis of correlation matrices. In R. B. Cattell (Ed.), Handbook Of multivariate experi- ‘mental psychology. Chicago: Rand McNally, 1966, pp. 438-458. Guttman, L., & Foa, U. G. Social contact and an intergroup attitude. Public Opinion gpartegly, 1951, 15, 45-53. Gynther, M. D., & Brilliant, P. J. Psychopathology and attitudes toward mental illness. Archives of General Psychiatry, 1964, 11, 48-52. Gynther, M. D., Reznikoff, M., & Fishman, M. Attitudes of psychiatric patients toward treatment, psychiatrists and mental hospitals. Journal of Nervous and Mental Disease, 1963, 136, 68-71. Hamersma, R. J. Construction of an attitude-behavior scale of Negroes and Whites toward each other using Guttman facet design and analysis. Unpublished doctoral dissertation. Michigan State University, 1969. Hamersma, R. J., & Jordan, J. E. Attitude-Behavior Scale: BW/WN—G. East Lansing, Mich.: Michigan State University, 1969, available from authors. 171 Harrelson, L. E. A Guttman facet analysis of attitudes toward the mentally retarded in the Federal Republic of Germany: Content, structure, and determinants. Unpublished doctoral diSsertation. Michigan State University, 1970. Hayman, M. Current attitudes toward alcoholism of psychi- atrists in Southern California. American Journal of Psychiatry, 1956, 112, 485-493. Hoyt, C. J. Test reliability estimated by analysis of variance. In W. Mehrens and R. Ebel (Eds.), Principles of educational and psychological measure- ment. Chicago: Rand, McNally, 1967, pp. 108—115. Jordan, J. E. Attitude-Behavior Scale--MR (ABS-MR). East Lansing, Mich.: Michigan State University, 1967, available from author. Jordan, J. E. Attitudes toward education and physically disabledlpersons in eleven nations. East Lansing, Mich.: Latin American Studies Center, Michigan State University, 1968. (a) Jordan, J. E. A Guttman facet theory analysis of attitudes and behaviors of blacks(negroes) and whites toward each other. East Lansing, Mich.: Michigan State University, 1968, available from author. (b) Jordan, J. E. Guttman facet design and development of a cross-cultural attitude toward mentally retarded persons scale. East Lansing, Mich.: Michigan State University, 1969, available from author. Jordan, J. E., & Hamersma, R. J. Attitude-behavior scale: black/white (ABS:BW and ABS:WN. U. S. 112268 version. East Lansing, Mich.: Michigan State University, 1969, available from authors (seven scales). Kahn, R. L., & Cannell, C. F. The dynamics of interviewing. New York: Wiley, 1957. ‘— Kaiser, H. F. Scaling a simplex. Psychometrika, 1962, 27, 155-162. Kastenbaum, R., & Durkee, N. Young people view old age. In R. Kastenbaum (Ed.), New thoughts on old age. New York: Springer, 1964. Lindquist, E. F. Design and analysis of experiments in psychology and education. Boston: Houghton Mifflin, 1953. 172 Lingoes, J. C. An IBM 7090 program for the Guttman-Lingoes multidimensional scalogram analysis - I. Behavioral Science, 1966, 11, 76-78. Ludwig, A. M., & Farrelly, F. The code of chronicity. Archives of General Psychiatpy, 1966, 15, 562-568. Maierle, J. P. An application of Guttman facet analysis to attitude scale construction: A methodological study. Unpublished doctoral dissertation. Michigan State University, 1969. Maierle, J. P., Jordan, J. E., & Whitman, R. M. Attitude Behavior Scale-ABS-MI. East Lansing, Mich.: Michigan State University, 1970. Manis, M., Houts, P. S., & Blake, J. B. Beliefs about mental illness as a function of psychiatric status and psychiatric hospitalization. Journal of Abnormal and Social Psychology, 1963, 67, 226-233. Mechanic, D. Relevance of group atmosphere and attitudes for the rehabilitation of alcoholics. Quarterly Journal of Studies on Alcohol, 1961, 22, 634-645. Merrill, 8. E., & Gunter, L. M. A study of patient atti- tudes toward older people. Geriatrics, 1969, 24, 107-112. Mindlin, D. F. Attitudes toward alcoholism and towards self. Qparterly Journal of Studies on Alcohol, 1964,' 25, 136-141. Morin, K. N. Attitudes of Texas Mexican-Americans toward mental retardation: .A Guttman facet analysis. Unpublished doctoral dissertation. Michigan State University, 1969. Mulford, H. A., & Miller, D. E. Measuring public acceptance of the alcoholic as a sick person. Quarterly Journal of Studies on Alcohol, 1964, 25, 314-323. Newcomb, T. W. Social psychology. New York: Dryden Press, 1954. Nunnally, J. C. Popular conceptions of mental health. New York: Holt, Rinehart and Winston, 1961. Riley, J. W., & Marden, C. F. The medical profession and the problem of alcoholism. Quarterly Journal of Studies on Alcohol, 1946, 7, 240-270. 173 Rosencranz, H. A. A factor analysis of attitudes toward the aged. Gerontologist, 1969, 9, 55-59. Ruble, W. L., Kiel, D. F., & Rafter, M. E. Calculation of least squares (regression) problems on the LS routine. Stat. Series Description No. 7, Agricultural Experi- ment Station, Michigan State University, 1966. (a) Ruble, W. L., Kiel, D. F., & Rafter, M. B. One way analysis of variance with unequal number of rpplications per- mitted (UNEQl routine). Stat. Series Description No. 13, Agricultural Experiment Station, Michigan State University, 1966. (b) Ruble, W. L., & Rafter, M. E. Calculation of basic statis- tics when missing data is involved (The MD-STAT routine). Stat. Series Description No. 6, Agricul- tural Experiment Stations, Michigan State University, 1966. Sapir, J. V. Social work and alcoholism. Annals of the American Academy of Political and Social Science, 1958, 315, 125-132. Shaw, M. E., & Wright, J. M. Scales for measurement of attitudes. New York: McGraw-Hill, 1967. Simon, A. The geriatric mentally ill. Gerontologist, 1968, 8, 7-15. Souelem, Omneya. Mental patients' attitudes toward mental hospitals. Journal of Clinical ngchology, 1955, 11, 181-185. Strassburger, F., & Strassburger, Z. Measurement of attitudes toward alcohol and their relation to personality variables. Journal of Consulting Psychology, 1965, 29, 440—445. Straus, R. Community surveys, their aims and techniques. Quarterly Journal of Studies on Alcohol, 1952, 13, 254-270. Suchman, E. A. The intensity component in attitude and Opinion research. In S. A. Stouffer (Ed.), Measure- ment and prediction. Princeton: Princeton Univer- sity Press, 1950. Swanson, R. M., & Spitzer, S. P. Stigma and the psychiatric patient career. Journal of Health and Social Behavior, 1970, 11, 44-51. 174 Tuckman, J., & Lorge, I. The effect of institutionaliza- tion on attitudes toward old people. Journal of Abnormal and Social Psychology, 1952, 47, 337-344. Tuckman, J., & Lorge, I. Attitudes toward old peOple. Journal of Social Psychology, 1953, 37, 249-260. Tuckman, J., & Lorge, I. The influence of changed direc- tions on stereotypes about aging. Education and Psychological Measurement, 1954, 14, 128-132. Tuckman, J., & Lorge, I. Attitude toward aging of indi- viduals with experiences with the aged. Journal of Genetic Psychology, 1958, 92, 199-204. Tuckman, J., Lorge, I., & Spooner, G. A. The effect of family environment on attitudes toward old peOple and the older worker. Journal of Social Psychology, 1953, 38, 207-218. " Winer, B. J. Structural principles in experimental design. New York: McGraw-Hill, 1962. Wolf, R. M. Construction of descriptive and attitude scales. In T. Husen (Ed.), International study of achievement in mathematics. New York: Wiley, 1967. Wolfensberger, W. P. Attitudes of alcoholics toward mental hospitals. guarterly Journal of Studies on Alcohol, 1958, 19, 447-451. APPENDICES [,4 \! U1 APPENDIX A ATTITUDE BEHAVIOR SCALE: ABS-MI 176 L.hjl1illl‘n llllllllllllll ATTITUDE E'HAVICR SCAI ll] [‘1 - ASS-MI DIRECTIONS This booklet contains statements of how people behave in certain situations or feel about certain things. You, yourself. or other persons often behave in the same way toward everyone, including mentally 111 persons. By mentally 111 persons we mean those children or adults whose behaviors, feelings, or emotions cause them to have difficulties with everyday problems which they are unable to solve without help. You also have some general ideas about yourself, about other persons like you, and about mentally 111 persons. Sometimes you feel or behave the same way toward everyone and sometimes you feel or behave differently toward the mentally 111. Here is a sample question: Sample l 1. Other people believe they are more attractive than most mentally ill persons. g) agree . uncertain 3. disagree . If others believe that mentally 111 persons have less chance than they have to be attractive. you should circle the number 1 as shown above, or if you are using an IBM answer sheet make a heavy dark line on the answer sheet between the two lines after the number as follows: 1. noon: 2. 22:22 3- 2:22: 4. 222:: 5. 222:: After each statement there will also be a question asking you to state how certain or sure you were of your answer. If you choose answer number 1 as above and were very sure your complete answer would be as follows: Sample g 1. Other people believe they H 2. How sure are you of are more attractive than this answer? most mentally ill persons. © agree 1. not sure . uncertain g. fairly 3. disagree very sure 1° loss 2 2:22 3 22:: 4 2:2: 5 22:: l 222 2 2:: 3 III-4 2:: 5 2:2 *a**a***** DO NOT PUT YOUR NAME ON THE BOOKLET ***********§**§***§ by: J. Paul Maierle John E. Jordan Coll ge of Education Mich gan State University Revised by: 112270-Rw Ross Whitman (Ix) ASS-I-MI Directions: Section 1 This section contains statements about ideas which other people have about mentally 111 persons. sheet number that indicates how others comparevthemselves to the mentally ill. Please answer all questions. Circle or fill in the answer Other people believe the following things about mentally 111 persons as compared to those who are not mentally ill. 1. Other people believe theiré—9 2. marriages are happier than those of most mentally 111 persons. 1. agree 2. uncertain 3. disagree Other people believe they are more intelligent than most mentally 111 persons. 1. agree 2. uncertain 3. disagree Other people believe they are more understanding than most mentally ill persons. 1. agree 2. uncertain 3. disagree Other people believe they are ess careless than most mentally 111 persons when invited to someone's home. 1. agree 2. uncertain 3. disagree Cther peOple believe they are more friendly than most mentally ill persons. 1. agree uncertain 2. 3. disagree 112270-RW 10. How sure are you answer? 1 2 3. very sure How sure are you answer? 1. not sure 2. fairly sure 3. very sure How sure are you answer? 1. not sure 2. fairly sure 3. very sure How sure are you answer? 1. not sure 2. fairly sure 3. very sure How sure are you answer? 1. not sure 2. fairly sure 3. very sure not sure fairly sure of this of this of this of this of this k0 ABS-I-HI Other people believe the following thinrs about mentally 111 persons as compared to those who are not mentally ill: ll. Other people believe their eating habits are better than those of most mentally ill persons. 1. agree 2. uncertain 3. disagree 13. Other people believe they are more helpful than most mentally ill persons are. 1. agree 2. uncertain 3. disagree 15. Other peOple believe that when they lend things to thers they are more reliable than most mentally ill persons. 1. agree 2. uncertain 3. disagree 112270-RW 12. 14. 16. How sure are you of this answer? 1. not sure 2. fairly sure 3. very sure How sure are you of this answer? 1. not sure 2. fairly sure 3. very sure How sure are you of this answer? 1. not sure 2. fairly sure 3.’ very sure -4- 50° II-MI - ‘g—ILJ Directions: Section II This section ill persons. contains statements about interacting with mentally Please choose the answer that indicates what you think others believe about interacting with mentally 111 persons. Most peOple believe the following about interacting with mentally ill persons: 17. 19. 21. 23. 25. Most people believe that others just like themselves are married to mentally 111 persons. 1. disagree 2. uncertain 3. agree Most people believe that others like themselves intellectually enjoy mentally 111 persons. . disagree . uncertain . agree KANl-J Most people believe that others like themselves relate understandingly to mentally ill persons. 1. disagree 2. uncertain 3. agree Most people believe that others like themselves invite mentally ill persons into their homes. 1: disagree 2. uncertain 3. agree Kost people believe that others like themselves have friends who are mentally ill. 1. disagree 2. uncertain 3. agree 112270—Rw 18. 20. 22. 24. 26. How sure are you of this answer? 1. not sure 2. fairly sure 3. very sure How sure are you answer? 1. not sure 2. fairly sure 3. very sure How sure are you answer? 1. not sure 2. fairly sure 3. very sure How sure are you answer? 1. not sure 2. fairly sure 3. very sure How sure are you answer? 1. not sure 2. fairly sure 3. very sure of this of this of this of this ‘c- J AES-II-MI Most people believe the following about interacting with mentally ill persons: 27. lost people believe that 28. others like themselves eat with persons who are mentally ill 0 1. disagree 2. uncertain 3. agree 29. Most people believe that 30. others like themselves accept help from mentally ill persons. 1. disagree 2. uncertain 3. agree 31. Most people believe that 32. others like themselves lend things to mentally ill persons. 1. disagree 2. uncertain 3. agree 112270-Rw How sure are you of this answer? 1. not sure 2. fairly sure 3. very sure How sure are you of this answer? 1. not sure 2. fairly sure 3. very sure How sure are you of this answer? 1. not sure 2. fairly sure 3. very sure I Ch I ’ S-III-NI i‘n Directions: Section III This section contains statements about ways in which you, yourself, should act toward mentally ill persons. Please choose the answer that indicates how you feel you should act or believe. In reSpect to mentally ill persons, do youlnyourselfl believe that it is usually right or usually wrong: 33. I should be willing to 34. How sure are you of this marry persons who are answer? mentally ill. 1. disagree 1. not sure 2. uncertain 2. fairly sure 3. agree 3. very sure 35. I should be willing to 36. How sure are you of this intellectually enjoy answer? mentally ill persons. 1. disagree 1. not sure 2. uncertain 2. fairly sure 3. agree 3. very sure 37. I should be willing to 38. How sure are you of this understand mentally ill answer? persons. ' 1. disagree 1. not sure 2. uncertain 2. fairly sure 3. agree 3. very sure 39. I should be willing to 40. How sure are you of this invite mentally ill answer? persons to my home. 1. disagree 1. not sure 2. uncertain 2. fairly sure 3. agree 3. very sure 41. I should be willing to 42. How sure are you of this be friends with mentally answer? ill persons. 1. disagree 1. not sure 2. uncertain 2. fairly sure 3. cg ee 3. very sure #3. I should be willing to 44. How sure are you of this eat with mentally ill answer? 1. disagree 1. not sure 2. uncertain 2. fairly sure 3. agree 3. very sure 112270-3' m I H H H I "4 H V‘ A“ -7, In respect to mentally ill persons, do youJ yourselfl believe that it is usually right or usually wrong: 45. I should be willing to 46. accept help from mentally ill persons. 1. disagree 2. uncertain 3. agree 47. I should be willing to 48. lend things to mentally ill persons. 1. disagree 2. uncertain 3. agree 112270-BW How sure are you of this answer? 1. not sure 2. fairly sure 3. very sure How sure are you of this answer? 1. not sure 2. fairly sure 3. very sure Directions: -Q- U ABS-IV-RI Section IV This act toward mentally ill persons. section indicates how you think you would act. answer that In respect to a mentally ill person would you, yourself: 49. 51. 53. 57. I woul marry someone who is mentally ill. 1. disagree 2. uncertain 3. agree I would intellectually enjoy a mentally ill person. 1. disagree 2. uncertain 3. agree I m would understand a entally ill person. I. disagree 2. uncertain 3. agree I i would invite a mentally ll person to my home. . disagree . uncertain would be friends with entally ill persons. 1. disagree 2. urcertain 3. agree I would eat ultq mentally ill persons. 1. disagree 2. unce tain 112270-RW 50. 52. 54. 58. 60. How sure are you answer? 1. not sure 2. fairly sure 3. very sure How sure are you answer? not sure fairly sure very sure DJIUP-J low sure are you answer? 1. not sure 2. fairly sure 3. very sure How sure are you answer? 1. not sure 2. fairly sure 3. very sure How sure are you answer? not sure fairly sure . very sure \nhak’ How sure are you answer? 1. not sure 2. fairly sure 3. very sure of of of of of of contains statements about how you think you would Choose the this this this this this this -9- ABS-IV-MI In respect to a mentally ill person, would you, yourself: 61. I would accept help from 62. mentally ill persons. 1. disagree 2. uncertain 3. agree 63. I would lend things to 64. mentally ill persons. 1 disagree 2. uncertain 3. agree 112270-RW How sure are you of this answer? 1. not sure 2. fairly sure 3. very sure How sure are you of this answer? 1.2 not sure 2. fairly sure 3. very sure Directions: -10- A3S-V-MI Section V This section the the mentally ill. following statements. How do you actually feel toward mentally ill persons: 65. 67. 71. 73- I feel unhappy about marrying a mentally ill person. 1. agree 2. uncertain 3. disagree I feel intellectually attracted to mentally ill persons. 1. disagree 2. uncertain 3. agree I am frightened about understanding a mentally ill person. I. agree 2. uncertain 3. disagree I am happy about inviting mentally ill persons to my home. . disagree . uncertain . agree KnNr-J I feel friendly toward mentally ill persons. 1. disagree 2. uncertain 3. agree I feel at ease about eating with mentally ill persons. 1. disagree 2. uncertain 3. agree 112270-RW 66. 68. 70. 72. 76. How sure are you answer? 1. not sure 2. fairly sure 3. very sure How sure are you answer? . not sure . fairly sure . very sure \J) {\JI-J How sure are you answer? 1. not sure 2. fairly sure 3. very sure How sure are you answer? 1. not sure 2. fairly sure 3. very sure how sure are you answer? ' 1. not sure 2. fairly sure 3. very sure How sure are you answer? 1 not sure 2. fairly sure 3. very sure of of of of of concerns actual feelings that people may have about You are asked to indicate how you feel about this this this this this this F4 -11- ABS-V-MI How do you actually feel toward mentally ill persons: 77. I feel all right about accepting help from mentally ill persons. 1. disagree 2. uncertain 3. agree 79. I feel tense about lending things to mentally ill persons. 1. agree 2. uncertain 3. disagree 12270-RW 78. 80. How sure are you of this answer? 1. not sure 2. fairly sure 3. very sure How sure are you of this answer? ‘ 1. not sure 2. fairly sure 3. very sure -12- A88-VI-NI Directions: Section VI This section concerns actual ergeriences you have had with mentally ill persons. Try to answer the following questions from the knowledge of your actual experiences. Experiences or contacts with mentally ill persons: Bl. I am married to a mentally 82. How sure are you of this ill person. answer? 1. disagree 1. not sure 2. uncertain 2. fairly sure 3. agree 3. very sure 83. I have intellectually 84. low sure are you of this enjoyed mentally ill answer? persons. 1. disagree 1. not sure 2. uncertain 2. fairly sure 3. agree 3. very sure 85. I have understood mentally 86. How sure are you of this ill persons. answer? 1. disagree l. not sure 2. uncertain 2. fairly sure 3. agree 3. very sure 87. I have invited mentally ill 88. How sure are you of this persons to where I live. answer? 1. disagree 1. not sure 2. uncertain 2. fairly sure 3. agree 3. very sure 89. I have friends who are 90. low sure are you of this mentally ill. answer? 1. disagree 1. not sure 2. uncertain 2. fairly sure 3. agree 3. very sure 91. I have eaten with persons 92. How sure are you of this who are mentally i 1. answer? 1. disagree I. not sure 2. uncertain 2. fairly sure 3. agree 3. very sure 112270-RN -13- ABS-VI-KI Experiences or contacts with mentally illg;ersons: 93. 94. I have accepted help from mentally ill persons. . disagree . uncertain 3. agree )FQH 96. I have loaned things to mentally ill persons. 1. disagree 2. uncertain 3. agree 112270-RW How sure are you of this answer? 1. not sure 2. fairly sure 3. very sure low sure are you of this answer? 1. not sure 2. fairly sure 3. very sure -1u- -. This part of the booklet deals with many things. Part of the questionnaire has to do with personal information about you. Since the questionnaire is completely anonymous or confidential, you may answer all of the questions freely without any concern about being identified. It is important to obtain your answer to eVery Question. Read each question carefully and do not omit any questions. Please answer b, circling the answer you choose or if you are Y i_ . using an IBM answer sheet, make a heavy dark line on the answer sheet between the two lines after the number you select. 97. Please indicate your age as follows: . 30 years of age or under 31-40 years KANH . Ql-5O years 4. 5l-6O years 5. 60 years or over 6. Write your exact age 98. About how much education do you have? 1. 6 years of school or less 2. Between 7 and 9 years of school 3. Between 10 and 12 years of school 8. Some college or university 5. A college or university degree 6. Write the number of years you went to school 112270-RW fix. y\)° p:- d ‘H ‘J P l 2. \Q o r: Kn -15- out how important is your religion to you in your aily life? Not at all important Not very ihportant Neither important nor unimportant Fairly itportant Very important 100. In respect to your religion, about to what extent do you observe the rules and regulations of your religion? 1. Almost never Rarely Occasionally Frequently Alnost always 101. Some people are nore set in their ways than others. How wou d you rate 102 0 E‘Jhr’at i l. 2. 3. 1+. C .2 U) yourself? I find it very difficult to change I find it 811 ghtly difficult to change I find it neither difficult nor easy to change I find it sciewhat easy to change I find it very easy to change your feeling about the following statement? "New methods of raising children should be tried out." 112270-RW 1. t\) K.) Almost always -16- 103. What is your feeling about a married couple practicing birth control? 1. It is altost never right 2. It is rarely right 3. It is occasionally right a. It is frequently right 5. It is almost always right 104. How long have you been hospitalized? l. 15 or more years . 10—15 years . 5-10 years 2 3 4. 1-5 years 5. Less than one year 6 . Write the number of years that you have been hospitalized 105. I feel responsible for my actions as follows: (circle one answer only) 1. Almost never 2. Barely 3. Occasionally 4. Frequently 5. Almost always 106. I like to be involved in some kind of work, recreational, or hobby activities as follows: (circle one answer only) 1. Almost never 2. Barely 3 Occasionally a. Frequently 5 . Almost always 112270-RW -17- 107. Privileges granted by a psychiatric institution such as a privilege card, passes, etc., should be earned: (circle one answer only) 1. Almost never 2. Rarely 3. Occasionally 4. Frequently 5 . Almost always ll2270-RW Q- U _‘| .1. This part of the quest‘onnaire deals with your exper- iences or contacts with uentally ill persons. Perhaps you have had much contact with mentally ill persons, or you may have studied about them. On the other hand, you may have had little or no contact with mentally ill persons, and may have never thought much about them at all. In the following questions, if more than one experience applies, please choose the answer with the highest number. 108. Have you had any experience with mentally ill persons? Considering all of the times you have talked, worked, or in some other way had personal contact with mentally ill persons, about how many times has it been altogether? 1. Less than 10 occasions 2. Between 10 and 50 occasions 3. Between 50 and 100 occasions a. Between ICC and 500 occasions 5. Kore than 500 occasions. 109. When you have been in contact with mentally ill persons, how easy for you, in general, would it have been to have avoided being with these mentally ill persons? 1. I could not avoid the contact 2. I could generally have avoided these personal contacts only at great cost or difficulty 3. I could generally have avoided these personal contacts only with considerable difficulty 4. I could generally have avoided these personal contacts but with some inconvenience 5. I could generally have avoided these personal contacts without an" difficulty or inconvenience 112270-Rw -19- 110. How have you generally felt about your experiences or contacts with mentally ill persons? I definitely disliked it I did not like it very much I neither liked it nor disliked it I liked it somewhat I definitely enjoyed it 111. In your contact or experience with mentally 111 persons, what opportunities do you have to associate with someone else such as friends, relatives, hospital personnel, etc. that are acceptable to you? 1. 2. 112270-Rw No one else is available Other people available are acceptable to me. Other people available are acceptable to me Other people available are acceptable to me Other people available are acceptable to me not at all not quite slightly fully LIFE This section of the booklet deals with how people feel about several aspects of life or life situations. Please indicate how you feel about each situation by circling the answer you choose, or if you are usinq an IBM answer sheet fill in the space after the number you select. 112. 118. ’41 m C) It should be possible to (l_9.11°, eliminate war once and for all. 1. strongly disagree 2. disagree 3. agree 4. strongly agree Success depends to a large 115. part on luck and fate. 1. strongly agree 2. agree 3. disagree a. strongly disagree Some day more of the 117. mysteries of the world will be revealed by science. 1. strongly disagree 2. disagree 3. agree 4. strongly agree By improving industrial and 119. agricultural nethods, poverty can be elirinated in the world. 1. stronrly disagree 2. ‘isagree 3. agree 4. strongly agree With increased medical 121. knowledge it should be ,. h: j ‘.-.~ ' possible to lengthen the axerage life Spin to 100 years or more. 1. strongly disagree 2. disagree 3. agree b. strongly agree 112270-RW How sure do you feel about your answer? not sure at all not very sure fairly sure . very sure twmw 0 How sure do you feel about your answer? 1. not sure at all 2. not very sure 3. fairly sure a. very sure How sure do you feel about your answer? not sure at all not very sure fairly sure . very sure (Timmy—J 0 How sure do you feel about your answer? not sure at all not very sure fairly sure . very sure Cxpnaw 0 How sure do you feel about your answer? not sure at all not very sure fairly sure very sure {1'me O O O lZQ. 126. 128. Someday the deserts wi be converted into good farming land by the a cation of engineering science. strongly disagree disagree agree strongly agree (3’me O O 0 Education can only help 125. people develop their natural abilities; it cannot change people in any fundamental way. 1. strongly agree 2. agree 3. disagree a. strongly disagree With hard work anyone can succeed. 1270 l. stronily disagree 2. disagree 3. agree U. strongly agree Almost every present 129. human oroblen will be solved in the future. strongly disagree disagree agree strongly agree CWbtv14 O 11227O-RW How sure do you feel about your answer? not sure at all . not very sure . fairly sure . very sure {:wl‘JP-J How sure do you feel about your answer? not sure at all not very sure fairly sure . very sure €Wgrvr4 How sure do you feel about your answer? 1. not sure at all 2. not very sure 3. fairly sure n. very sure How sure do you feel about your answer? not sure at all not very sure fairly sure very sure CKANl—J ~c2- " *nv'h Y ‘1 :‘CO L. .v-\‘ s A‘I! TI LJ'.“ t_. 3 section of the uestionnaire deals with information about q Please circle your answer, or if you are using t fill'In the space after the number you select. 1. strongly disagree 2. disagree 3. agree 4. strontly agree 131. Psychiatrists try to teach mental patients to hold in their strong emotions. 1. strongly agree 2. agree 3. disagree 4. strongly disagree 2 Fost of the rentally ill cases are found in peOple over fifty years of age. F“ \i) o I. strongly agree ‘ 2. agree 3. disagree 4. strongly disagree 133. Nany of the reotle who go to mental hospitals are so e to return to work in society again. strongl' disagree disagree agree strongly agree chter O O 0 13h. X-rays of the head will not tell whether a person is likely to develop mental illness. 1. strongly disagree . disagree . agree . strongly agree Pinto fiost peerle can recognize the type of perSon who is likely to have a nervous breakdown. F.) K!) \n O . strongl agree . agree . disagree . stronély disagree -C\JPJFJ 112270-Rw -23- 136. People who are likely to have a nervous breakdown pay little attedtion to their personal appearance. 1. strongly agree 2. a.;ree 3. disagr ree 4. strongly disagree 137. Almost any disease that attacks the nervous system is likely to bring on mental illness. . strongly agree agree disagree strongly disagree Cwbrv+4 138. Fost people who "go crazy" try to kill themselves. . strongly agree agree disagree strongly disagree FWNH 139. If a child is jealous of a younger brother, it is best not to let him show it in any way. 1. strongly agree 2. agree 3 disagree 4. strongly disagree 140. A person cannot rid himself of unpleasant memories by tryingh nard to forget them. 1. strongly disagree 2. disagree 3. agree 4. strongly agree 141. {el ng t‘ne mentally ill rerson with his financial and social problezs will not cure his disorder. 1. strongly disagree 2. disagree 3. agree 4. strongly agree 142. float of the reop le in .nental hospitals speak in words that can be understood. 1. strongly disagree 2. disagree 3. agree 4 strongly' agree 112270-RW 1&3. lad. 1&6. The following four questions are Early training will not faster. 1. strongly 2. disagree 3. agree a. strongly Older reorle have from a nervous 1. strongly 2. aaree 3. disagree 4. strongly There is not much develops a mental 1. strongly 2. agree 3. disagree 4. strongly Few of the people treattent. «tuJNrJ all of them. 147. l. 2. l. 112270-RW strongly agree disagree strongly Female. Mble Karried Single Divorced Widowed Separated -24- make the child's brain grow disagree agree a less difficult time recovering breakdown. agree disagree that can be done for a person who disorder. agree disagree who seek psychiatric help need the agree disagree general in nature. Please answer Ilease indicate your sex. is your marital status? -25- 1&9. what is your religion? 1. I prefer not to answer 2. Catholic 3. Protestant a. Jewish 5. Other or none 150. My therapy has consisted mostly of the following: (Circle one answer only) 112270-RW l. 2. I have not had any therapy *2 .edications Assignment to activities such as Occupational Therapy (O.T.), Corrective Therapy (C.T.), etc. Group therapy Individual therapy APPENDIX B STATISTICAL MATERIAL 202 2(33 TABLE 18.--N's, means, and standard deviations for psychiatric and normal samples on 12 dependent1 and 18 independent variables. 3 ssn4 58A5 117537 5167 variable Y so Y so 3? so Y so Attitude Content Stereotype 13.1 2.9 11.1 2.6 15.2 4.8 9.9 2.4 Normative 13.8 3.3 14.7 4.2 17.9 3.9 15.2 4.6 Moral Evaluation 19.9 3.0 18.1 3.1 18.3 3.8 16.9 4.3 Hypothetical Action 19.2 3.6 17.0 3.1 16.9 3.5 17.5 4.6 Personal Feeling 17.9 3.1 15.7 3.0 17.6 3.3 16.5 3.9 Personal Action 17.7 4.0 16.2 3.2 13.8 4.3 16.8 3.9 Attitude Intensity Stereotype 16.8 3.5 19.9 2.9 19.1 3.5 20.8 4.2 Normative 16.7 4.0 19.1 3.0 19.7 3.7 21.0 4.5 Moral Evaluation 21.0 2.5 20.4 2.8 19.9 3.5 22.1 3.8 Hypothetical Action 20.1 3.1 19.4 3.8 19.7 3.5 22.0 3.6 Personal Feelings 20.2 3.2 19.2 3.6 21.0 3.4 22.0 3.6 Personal Action 20.8 2.7 20.6 2.7 19.9 3.5 22.4 3.2 Demographic Age 1.1 .3 2 9 .8 2.6 1.0 5 O .1 Education 4.9 .3 2 8 .8 3.0 .9 2 4 1.1 Relig.-Importance 3.2 1.5 3 7 1.0 3.8 1.3 3.6 1.4 Relig.-Adherence 3.2 1.4 3 5 1.1 3.7 1.3 3.8 1.2 Hospitalization 4.9 .6 4 9 .4 2.9 1.4 2.6 1.4 Change Orientation Self Change 2.9 1.1 2.6 1.1 2.6 1.3 2 9 1.6 Child Rearing 3.8 .9 3.5 1.1 3.3 1 3 3 2 1.3 Birth Control 4.7 .7 3.9 1.2 3.4 1.5 3 4 1.5 ChronicityVOrientation Responsibility 4.8 .7 4.5 .8 4.5 1.0 4.6 .9 Work Involvement 4.3 1.0 4.0 1.2 3.7 1.3 3.9 1.3 Privileges 3.4 1.1 4.3 1.1 4.0 1.4 4.5 .8 Contact Amount 2.4 1.4 2.4 1.4 3.4 1.6 3.7 1.6 Avoidance 3.0 1.6 3.7 1.6 2.6 1.7 2.5 1.6 Enjoyment 3.7 .9 3.2 .9 3.1 1.1 2.9 1.2 Alternatives 4.4 1.2 4.2 1.1 3.9 1.4 4.1 1.3 Value Efficacy—Content 24.1 2.7 24 6 3.2 24.0 4 O 25.9 3.4 Efficacy-Intensity 29.3 4.6 30.8 3.1 30. 4 9 31.9 5.2 MI Knowledge 52.6 6.1 47.2 4.2 46.1 5.9 45.8 6.1 1Attitude content and intensity variables SAlcoholics 2Demographic and all subsequent variables 6Schizophrenics 3Variables are described in Chapter III 7Geriatrics 4 Normals 204 mHsEMoz v moauumauoub HHH woodman ca nonwuomoo mus moansaum>m moacmuzmouanomm ona mam>oa Hznmmam mowaozooadm homowmmmH m.ma v.mm m.mH >.vm m.mH «.mm N.MH m.vm poswneoo muwmcoucH was usmuGOUuumusoammm ~.ma o.m¢ H.oH m.Hm n.~a m.ov m.ma w.ov coauod denounce m.mH m.av v.~H m.m¢ m.oa m.mm m.HH H.5V meadows HMGOmHom m.ha H.mv o.ma ~.mv m.~H a.mv m.ma a.~m coauod Hmowuocuommm «.ma «.mv m.ma m.mv n.ma m.he m.HH H.mm cowumsas>m Heuoz m.hH m.hm H.¢H H.h¢ m.mH v.vm v.HH m.am o>wuseuoz w.m H.mH H.5H m.>m m.m m.HN m.oH ~.m~ omauomumum |mmnwmfioo aumncoucH 0cm usoucounnopsuwuud mm M mm N am m cm M moansaus> mam omnaa mmmm vam .Umcwafioo auwmcmucw was pamHGOOIumoHnmaum> mucopcomoo m was unaccommpcw H :0 moamfimm asshoc was cauusanommd How meowumH>op pumocmum was .msmoE .m.mnu.ma mqmda H 205 hII:1 a moauuswuoum mo.v.m« muacmucmonwnome amum> ova> one acqusuaamuadmon 30H 0» ncommmuuoo mosflm> :mamw mowaocooadm m «EH0 moamzh a z~ mmamEomw onmm season can xom u0w poumaflp I I I I I I I I I Hm.H ma.v mo.v mo.v nm.m mH.v vm.v wo>auscumuad oAz .mAz .aAz .vq.s om.m mo.~ sm.~ ms.~ mm.~ se.m acossoflcm oAd .mA< .zA< amH.h am.N mm.m mn.m Hm.~ mm.m mo.m occupwo>< (A0 .ZAO .wum I I I I I I I I I mm.H nm.m mm m om.m mm.m Hm.m m~.v ucoso>ao>cu xuoz I I I I I I I I I mm. mm.v mh.v ms.v wm.v oo.v om.v xuwawnwmcommmm coaumusoauo smaoacouno mA< .oAz .mAz ..~ hm.m mm.v om.v meowumnwasuwamom 2A0 .zAm «cm.m Nm.m wh.m mm.m mw.m mm.m ma.m mucmumnp¢s.mwamm zAm .zA< ama.m vm.m m~.m mo.m mm.m vm.m ma.m mocmquQEHI.maaom oAm .scom .om»: oAm .UAZ .mAz .¢Az amm.m hm.mv om.wv ~H.Nv om.hv mm.mv Ho.mm H0902 zAm .tAm aom.v mm.wm wa.mm oq.nm mm.av m~.vm mm.Hm o>ausEuoz wq mamom HZImm< a I m “c4 m ue< m flea m flv< m he< m sea «gamma memo: 0H Haas: m IIIII moannaum> z . H mm whom mUHm vmhaa mdmm Nme Hafiuoc paw .oocflneoo xuwmcoucw use acoucoo HxImm< new momuauonImoamfism uauuwwsoamm sew umou memos onHuHsé was .m.m H .msmoe pmumsnpm .monwm mHmEsmII.o~ mqm