jalfitulfang-l ‘7’"".‘~‘-.".‘-‘u‘I‘I‘u.‘-t‘.-"'.s',". 7:33:44.‘I.\'.:;V..:‘j..._- T§I.,rv...:..-,-‘,,' ,-,n:.‘...- ‘, . I . v H . IT ‘, . ._ ‘ ’ A ‘_ {,5 .. u . THE DIFFUSION OF THE COMMUNITY LODGE Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY ROBERT N. HARRIS, JR. 1972 I . + 3 IIIIIIIIIIIIIIIIIIII 001 .‘ I .. . LIBRARY Michigan State University This is to certify that the thesis entitled The Diffusion of the Community Lodge presented by Robert N. Harrie, Jr. has been accepted towards fulfillment of the requirements for Ph.D. degree mm CagZW" -- V Major professor 2" magma av 1‘ IIMG 8| SIJNS' 800K HWY INC. LIBR551.3'.'.'.9.E§?. ABSTRACT THE DIFFUSION OF THE COMMUNITY LODGE By Robert N. Harris, Jr. This study investigated the diffusion throughout the U.S. mental health system of an innovation in the treatment of chronically hospitalized mental patients. This innovation, the community lodge, was developed by George W. Fairweather and his co-workers. Previous ex- perimental efforts had shown the need for the lodge, the efficacy of it and the variables important in implementa- tion of such a program. The present study is the logical extension of those research efforts. The following hypotheses were derived from the theoretical positions of four social change models; the ' research, development and diffusion perspective, the .problem-solver perspective, the socialninteraction per- spective and experimental social innovation. There is a significant positive relationship between progress towards adoption of the lodge and. l. the existence of specialized roles for the input of new programs, Robert N. Harris, Jr. 2. the hospital's goals of seeking in- formation about new programs, 3. change in leadership in the hospital, 4. the occurrence of a crisis in the hospital, 5 a broad pattern of decision-making in the hospital, 6 a "change orientation” in the hospital, 7. a "systemic" perspective in the hospital. There is a significant negative relationship be- tween progress towards adoption of the lodge and: 8. "local pride" expressed by the hospital. Measures of each of the variables to test these hypotheses were based upon telephone questionnaire re- sponses gained from 244 State and V.A. hospitals through- out the U.S. These included 102 hospitals which had pre— viously indicated no willingness to adopt the lodge (NO/NO hospitals), 117 hospitals which had previously indicated a willingness to receive a brochure, attend a workshop or develop a demonstration ward, but would not volunteer to implement the lodge (YES/NO hospitals), and 25 hospitals which previously indicated a willingness to implement the lodge (volunteers). Two other sources were also used: background data from the previous hospital implementation study and demographic information available from the Ameri- can Hospital Association journal, Hospitals. Taken together, the variables measured the diffusion of the lodge, hypothesis- testing information and innovative program descriptions. The results reveal little diffusion of the lodge in the NO/NO and YES/NO hospitals with significantly greater diffusion occurring in the original volunteers. This is Robert N. Harris, Jr. derived from the chi-square comparing differences among hospitals on degree of adaption of the lodge, which is significant beyond the .001 level. This finding is the most salient of the present study. Apparently diffusion must, in fact, be planned. Those hospitals which had been most active previously (in YES/NO and volunteer hos- pitals) did, in fact, diffuse more. This comparative analysis led to the use of three separate cluster analyses to test the hypotheses. The results of the preset cluster analyses did not support a single hypothesis as contributing to the diffusion of the lodge. There was some indication of the importance of a broad pattern of decision-making in the volunteer hospitals. Other results indicate that diffusion is unrelated to other domains except in the NO/NO hospitals. This in— dicates that with no intervention, diffusion is related to adoptiveness-innovativeness, expenses and superintendent influence, but that the intervention attempt itself alters other relationships which may exist without it. In addi- tion, the cluster analysis of the volunteer hospitals re- veals several interesting findings. The specialized roles for input of new programs is related to stability, not adoption of new programs. Such adoption seems to occur in those hospitals which place new program responsibility in less well-established hands. Robert N. Harris, Jr. Generally, the results of both comparative and correlative analyses reveal the following findings: 1. More active implementation attempts lead to greater diffusion of the lodge. 2. Diffusion of a complex social innovation is relatively unrelated to other organizational, attitudinal and demographic variables, and therefore; 3. There is little correlative evidence to sup- port any of the eight hypotheses. Several limitations of the present study were discussed. These included the lack of diffusion of a complex social change and its effect on our analyses, telephone rather than face-to-face interviewing tech- niques employed and the limited N size in the volunteer hospitals. Finally, recommendations for future research were made. These included researches to investigate methodological questions, differences between different types of innovations, the importance of an active change agent and the possibility of developing diffusion centers. THE DIFFUSION OF THE COMMUNITY LODGE By Robert NTwHarris, Jr. A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1972 ACKNOWLEDGMENTS After one final perusal of my dissertation, I fully realized the importance of those who helped make this possible. It would, of course, be impossible to thank them all, but I will try nonetheless. First I would like to thank my thesis chairman, Dr. George W. Fairweather. He gave me the chance to make my graduate training a meaninfgul experience and prepare for a career helping others in both a humani- tarian and_scientific way. His advice, support and in- terest made this dissertation a reality.‘ In addition, I owe a debt of gratitude to Dr. Louis Tornatzky for introducing me to Thome and Associates, the Janitorial Aces and other outstanding organizations. Also, I thank Dr. Tornatzky for help in the design, instrumentation and data analysis of this thesis. I must thank Dr. Dozier Thornton and Dr. Lawrence I. O'Kelly for their support and understanding throughout the writing of my disserta- tion. Next, I'd like to express thanks to Baron Perlman and Mike Denny for their great efforts in battling Bell Telephone and somehow keeping hospital administrators on the phone for forty minutes. Without their help, and the ii cooperation of all my hospital respondents, there would have been no thesis. I'd also like to thank all of the ecological psychology group for their comments and sup- port during this, the first thesis for our troops. Ms. Gudrun Gale and Ms. Kathy Looney are especially to be congratulated for putting up with telephone inconven- ience and requests for help during all parts of my work. Friends and family are the backbone of any effort the magnitude of a dissertation; for these people must bear with the author through both good and bad times. That they do, is a tribute to their stamina and love. Thank you Harvey, Leah, Dave and Marcia. My family has given me the support one can find nowhere else. I thank both my families for seeing me through to this moment. Finally, I dedicate this dissertation to my wife Fran for her love and confidence in me. In my most fer- vent "fourth grade math" moods, her strength carried me through. All I can say is thanks and I love you. iii TABLE OF CONTENTS Page ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . ii LIST OF TABLES . . . . . . . . . . . . . . . . . . vii LIST OF FIGURES . . . . . . . . . . . . . . . . . . viii INTRODUCTION . . . . . . . . . . . . . . . . . . . 1 Social Change Models . . . . . . . . . . . . . . 2 The Research, Development and Diffusion Perspective (R, DGD) . . . . 2 The Problem- Solver Perspective (P- S). . . 3 The Social- Interaction Perspective (8- I) . . . . . . 4 Experimental Social Innovation (E. S. I. ) S A Functional Description of E.S.I. . . 7 Background - From Traditional to Innovative Treatment of Mental Patients . . . . 7 Creating and Evaluating a New Model - The Community Lodge . . . . . . . . . . . . . . 9 Experimental Implementation . . . . . . . . . . 11 An E.S.I. Diffusion Study . . . . . . . . . . . . 12 Basic Concepts of Diffusion . . . . . . . . . . 13 , Hypotheses . . . . . . . . . . . . . . . . . . lo//// METHODS . . . . . . . . . . . . . . . . . . . . . . 23 Sampling Units . . . . . . . . . . . . . . . 23 Data Collection Procedures . . . . . . . . . . . 24 Measurement of Variables . . . . . . . . . . . . 28 Hospital Study Data . . . . . . . . . . . . . . Z8 Demographic Information . . . . . . . . . . . . 29 Phone Questionnaire . . . . . . . . . . . . . . 29 Data Analysis . . . . . . . . . . . . . . . . . 33 iv RESULTS . . . . . . . . . . . . . . . . . . Amount of Diffusion . . . . . Diffusion and Degree of Adoption- A Closer Look at the 25 Volunteer Hospitals . . . . . . . Cluster Analysis of Diffusion and Hypothesis- Testing Variables . . . . . . The NO/NO Hospitals . . . . . . . . . . The YES/NO Hospitals . . . . . . . The YES/YES and NO/YES Hospitals (Volunteers) Hypothesis-Testing Variables . . . . . . DISCUSSION Diffusion . Amount of Diffusion . . . Diffusion Hypotheses . . . Relationship Between Diffusion Cluster and Other Clusters Overview of Additional Results The Non- Volunteers . The Implementation Volunteer Hospitals . . .~. . Interpretations and Conclusions . Limitations . . . Planned and Un lanned Diffusion . Future Researc . . . . . . . . . BIBLIOGRAPHY APPENDICES APPENDIX A Phone Questionnaire Introduction APPENDIX B Questionnaire Items Used to Measure Degree of Adoption of the Community Lodge (Diffu) . . . . . . . . . Page 35 35 37 39 40 48 55 60 66 66 66 68 72 73 73 78 81 81 83 85 88 93 94 Page APPENDIX C Questionnaire Items Used to Measure Hypotheses . . . . . . . . . . . . . . . 97 APPENDIX D Diffusion-Adoption Stages . . . . . . . . . . . 101 APPENDIX E Community-locus Score . . . . . . . . . . . . . 104 APPENDIX F Autonomy Score . . . . . . . . . . . . . . . . . 105 vi Table 1. lo. 11. LIST OF TABLES Experimental Social Innovation and the Models of Dissemination and Utilization . . . . . . . . . Comparison of Amount of Diffusion Between Hospitals Indicating Differ- ential Willingness to Adopt the Lodge in Previous Research Efforts . . . . The Ten Clusters in the NO/NO Hospitals . Correlations Between Oblique Cluster Domains for the NO/NO Hospitals . . . The Nine Clusters in the YES/NO Hospitals . . . . . . . . . . . . . . . Correlations Between Oblique Cluster Domains for the YES/NO Hospitals . . The Three Clusters in the Implementation Volunteer Hospitals . . . . . . . . . . Correlations Between Oblique Cluster Domains for the Implementation Volun- teer Hospitals . . . . . . . . . . . . Cluster Loadings of Hypothesis-Testing Variables on Diffusion Score . . . . . Clustered Measures in the Implementation Non-Volunteer and Volunteer Hospitals . Mean Values of Amount of Influence - Implementation Non-Volunteers vii Page 36 41 43 49 51 56 58 61 74 78 LIST OF FIGURES Figure l. The Diffusion Curve 2. The Adoption Curve . . . . . 3. Comparison of Diffusion and Degree of Adoption for the 25 Volunteer Hospitals . . . 4. Amount of Diffusion in the NO/NO Hospitals . . . . . . viii Page 14 16 38 44 CHAPTER I INTRODUCTION This study investigates the diffusion throughout the U.S. mental health system of an innovation in the treatment of chronically hospitalized mental patients. George W. Fairweather and his co-workers developed the community lodge program for these patients after years of research and evaluation of both traditional and in- novative treatment programs (Fairweather, 33 31., 1960; Fairweather, 1964, 1967; Fairweather, gE_§1., 1969). The empirical evidence reported in Community Life for the Mentally Ill (Fairweather, Sanders, Cressler and Maynard, 1969) indicated that the lodge program de- served further efforts at implementation. These efforts were recently completed and will be reported in Changing Mental Hospitals (Fairweather, Sanders and Tornatzky, 1972). The present investigation is a logical exten- sion of previous research efforts, and will examine the effects of planned and unplanned social change. To place this study in perspeCtive, however, it is neces- sary to explore the diffusion process and its relation- ship to different social change models. Social Chaggg Models In Planning for Innovation through Dissemination and Utilization of Knowledgg, Havelock (1971) provides three conceptual frameworks for diffusion research. In his comprehensive review of nearly four thousand studies, Havelock develops three models of dissemination and util- ization (D8U)1; the research, development and diffusion perspective (R,D&D), the problem-solver perspective (P-S), and the social-interaction perspective (S-I). The re- search approach created by Fairweather (1967) and followed by empirical research in 1969 and 1972 enhance these mo- dels by developing a fourth model, Experimental Social Innovation (E.S.I.), which incorporates concepts from these three approaches and places them in an experimental perspective. The Research, Development and Diffusion Perspective (R,D&Dj Guba (1966) developed a typical model for this perspective. The phases described by Guba (1966) and Guba and Clark (cited in Havelock, 1971) include research, development, diffusion and adoption. Of the three models of dissemination and utilization, the R,D&D model views 1Dissemination and utilization is viewed as the "transfer of messages by various media between resource systems and users." (Havelock, 1971). the process of change from the earliest point in time, research. Research "may provide a basis for innovation if anyone else is clever enough to develop an applica- tion from it." (Guba, cited in Havelock, 1971.) Development includes invention and design (Guba, 1966), engineering and manufacturing (Havelock and Benne, 1967), or design and evaluation (Brickell, in Miles, 1964). It is development "which is at the heart of change." (Guba, 1966.) For Guba, diffusion includes both dissemination and demonstration. These activities "create widespread awareness of the invention among practitioners." (Guba, 1966.) Finally adoption includes trial, installation and institutionalization. The objective at this stage is the "incorporation of the invention into a function- ing system." (Havelock, 1971.) The R,D&D model, then, views change from an early point in time, and from the view of the originator and developer. These considera- tions, in particular, distinguish this model from the next tWO . The Problem-Solver Perspective (P-S) This model is based on the early work of Lewin (1951) who refers to three major stages of change, un- freezing, moving and freezing. Lippitt, Watson and‘ Westley (1958) expanded this basic model to include de- veloping the need for change, establishing a change re- lationship, diagnosis of client problems, establishing goals of action, transforming intention into actual change, generalizing and stablizing the change and achiev- ing a terminal relationship. Other authors concerned with this basic model include Mann and Williams (1960), Thelen, (in Watson, ed., 1967) and Miles and Lake (in Watson, ed., 1967). In the P-S model the receiver initiates the pro- cess of change by identifying a need. He actively searches for the innovation to solve his problem. Actual change usually involves outside assistance in implementing the change. As such, the P-S model is similar to the develop- mental aspect of the R,D&D model. The Social-Interaction Perspective (S-I) Rogers (1962, 1971) is most closely associated with this perspective. The focus of his model is on chan- nels of communication within the receiver group and with the stages through which individuals pass as they make progress towards adoption of innovations. These stages are knowledge, persuasion, decision and confirmation. Other authors concerned with this basic model are Wilkening (1962), Beal, Rogers and Bohlen (1957) in agriculture, and Coleman, g£_§l. (1966) in medical innovation. This model was derived largely from cases of "unplanned changes," whose development is characterized by the stages described above. Rogers (1962) cites sev- eral studies which support his concept of stages. In this manner, the 8-1 model assumes that both research and development have occurred, and that the major con- cern is for the spread of the innovation. Since the present study emphasizes the diffusion of a mental health innovation, this model will be of great signifi- cance to this paper. Experimental SocialFInnovation (E.S.I.) We have examined three models of dissemination and utilization research. A fourth model is now being developed which places the processes of innovation and implementation in an experimental framework. Histori- cally, survey research has most often been used as the basis for describing and conceptualizing different models of DGU. Fairweather (1967, 1972), in developing Experi- mental Social Innovation specifies the processes of inno- vation and implementation in conceptual stages that can be subjected to experimental techniques. Its common ele- ments are model-building, evaluation, implementation and diffusion; all in an experimental framework. The rela- tionship between these four models is indicated in Table I. Table 1. Experimental Social Innovation and the Models of Dissemination and Utilization. Models of Dissemination and Utilization Research, Experimental Development Problem- Social- Social and Diffusion Solver Interaction Innovation (R,D&D) (P-S) (S-I) (E.S.I.) Research Research Research Experimental Model—Building Assumed Assumed and Evaluation Definition Naturalism: Innovation COmparison Context Evaluation Responsibility Cross-disciplinary Development Need for Development Experimental change Implementation Assumed Establish relationship Approach Examine goals Persuasion Select alternatives Plan implemen- (Adoption) tation (Adoption) Installation Terminate re- lationship Knowledge Diffusion- Diffusion Diffusion (Adopt1on) Persuasion Decision (Adoption) Confirmation May include: approaching persuading A Functional Description of E.S.I. Backgppund - From Traditional to InnovatiVe Treatment of Mental Patients The innovative model presented in this study was developed against the background of many periods of change in the treatment of mental patients. Nicholas Hobbs (1969) cites three periods of change so radical that each is re- ferred to as a mental health revolution. The first revolu- tion is identified with Phillippe Pinel, William Tuke, Benjamin Rush and Dorthea Dix. These zealous reformers in- sisted that insane people be treated with kindness, rather than beatings. But in spite of their good intentions, the "cumulative effects of industrialization, massive urbaniza- tion, ... weakening of traditional family and local commun- ity ties...and many other social changes, led to a construc- tion boom in large, congregate, custodial institutions in France, England and the United States." (Pasamanick, g£_§l., 1967;) In these institutions "treatment" was custodial, rehabilitation non-existent and isolation from the outside world of prime concern. Such institutions give sound his- torical bases to the popular conceptions of mental health treatment models as forwarded by Goffman (1962), Kesey (1962) and others. The second revolution found its leader and inno- vator in Sigmund Freud and his preoccupation with the in- trapsychic life of man. Freud's disciples made individual psychoanalytic treatment the preferred approach to both chronic schizophrenics and middle class housewives. This innovative approach spread throughout the world and con- tinues to appeal to much of the mental health establish- ment. Such appeal grows despite numerous outcome studies which reveal little or no improvement due to psychotherapy (Eysenck, 1966). The "psychotherapeutic" revolution found itself subject to experimental work. In order to experimentally investigate the effects of such a model of treatment, Fairweather began a series of evaluative studies in 1955. The results of the first study, . . . showed that patients who participated in the three most common hospital treatment pro- .grams (models) [(1) individual psychotherapy, (2) group psychotherapy, and (3) living and working together as groups] do no better or worse in community adjustment 18 months after release from the hospital than those patients who had simply worked in the hospital setting (a fourth model). Fairweather, unpublished manuscript, 1972 In 1958 the traditional ward was further compared to a small group ward (another new model) in which patients were organized into problem solving groups. The findings indicated that such small groups could be formed and that autonomy and morale were enhanced. But patients still re- turned to the hospital as quickly as the more tradition- ally treated patients (Fairweather, 1964). The next model developed, the lodge, corresponds to the third revolution cited by Hobbs. In the third revolution, psychiatry "has shifted the major emphasis in mental health care out of the institution and back to the community." (Pasamanick, §£_§l,, 1967.) The commun- ity focus pervades recent innovative attempts in treating mental patients. These include half-way houses, foster care homes, independent living units, "enabler" programs, Home Care Projects (Pasamanick, op 31., 1967), Community Lodges (Fairweather, 33.31., 1969), etc. Such programs vary in the degree to which patients are autonomous, but all share the locus of the community. Creatin and Evaluatin a New 0 e - The Community 0 g§_ The transition of mental health programs from the hospital to the community requires the patient to be pre- pared for a new environment. The subordinate social sta- tus enforced in the hospital does not prepare the patient for his new situation. "Patients are administered to by the physician and professional staff, and their expecta- tions are organized around this lowly position." (Fairweather, g£_§1,, 1969.) Many "innovative" after- care programs retain such marginal status as part of the Community treatment. The patient is seen as a child, guided by house parents or foster parents. He is not al- lowed to develop as a first-class citizen until release and "cure." 10 In an attempt to develop first—class citizenship for the mental patient in a community setting, and to provide that person situational support, employment and extended tenure in the community, Fairweather and his co-workers created a new model of treatment, the commun- ity lodge. In an earlier study, Fairweather (1964) found that community tenure was intimately related to the patient in the community enjoying a socially suppor- tive living situation, frequent employment and employ- ment in a low-status job. These findings led to the development of the community lodge, which is fully des- cribed in Communipy Life for the Mentally Ill (Fairweather, g£_§1., 1969). Briefly, the lodge involves the training of groups ofpatients so that they may be prepared for full, first-class citizenship outside the hospital. After several months of small group training in the hospital, 10-15 patients were moved into their home in the community. In the community they had no live-in staff, a patient self-governing body, free exit and entry provisions, self-medication, and a business of their own to help support themselves. This social organization re- sulted in feelings that the patients were worthwhile, productive citizens. Evaluation of the program concluded that such a program is not only feasible, it is better than other existing in-hospital and post-hospital treat- ment facilities on several key dimensions. An evaluation 11 comparing the lodge members to a control group of in- hospital small group ward patients with traditional after- care facilities indicated that the lodge was significantly more effective in reducing recidivism and increasing em- ployment. In addition it was less expensive than the tra- ditional program. Expprimental_lmplementation Subsequent attempts to get the lodge adopted 10- cally and elsewhere revealed that presenting this research data was not sufficient to influence hospitals to accept lthe program. Therefore, Fairweather and his co-workers decided to attempt to nationwide experimental implementa- tion of the lodge program. Two hundred and fifty-five State and V.A. hospitals in the U.S. (virtually the en- tire population) were contacted about the lodge program. Several experimental conditions were evaluated. They were: 1. Method of presentation (brochure, workshop, development of a demonstration ward) 2. Status of initial contact (superintendent, psychiatry, psychology, social work, nursing) 3. State or V.A. hospital 4. Urban or rural hospital For the hospitals which volunteered to implement the lodge after initial persuasion attempts, a further 12 condition was tested. This involved the difference be- tween active, consultant help and written, manual help in developing the lodge. Preliminary analyses reveal that method of ap- proach presentation, implementation condition (written vs. action) and certain group process measures in the hospitals distinguish between hospitals which volunteer for and implement the lodge and those which refuse the attempts. Thus far it appears that the most active ap- proaches are the most successful. A detailed analysis of this study will appear in a forthcoming volume (Fairweather, Sanders and Tornatzky, 1972). An E.S.I. Diffusion Study This background information was necessary to understand clearly the study that constitutes the main concern of this paper. It is specifically concerned with the diffusion of the lodge program, i.e., what happened after implementation. Of the original 255 hospitals, only 25 formally volunteered to implement the lodge. There are therefore a population of 230 hospitals which were not subjected to implementation attempts by the research team. Some of these hospitals received a brochure, others attended a workshop, and still others set up a demonstration ward, but none of the 230 were part of any further planned implementation. 13 This condition raises several questions that need to be answered. Have the unconvinced hospitals adopted the lodge anyway? .What kinds of programs have these hos- pitals been adopting since the last contact with them? Have the 25 volunteer hospitals diffused the lodge con- cept? What organizational and attitudinal variables are important in the spread of this innovation? This study tries to answer these and other questions. Basic Concepts of Diffusion Discussion of diffusion is in each social change model described in Table 1. Now let us review the con- cepts of diffusion and adoption. Diffusion is the "process by which an innovation spreads among members of a social system." (Rogers, 1971.) An innovation is an idea, practice or object perceived as ggg_by the adopting population. For example, humane treat- ment of patients and individual psychotherapy were innova- tions at one time, and have since spread through the men- tal health system. The social system may be defined as any "collec- tivity of individuals, or units, who are functionally differentiated and engaged in collective problem-solving with respect to a common goal." (Rogers, 1971.) Thus, we may be referring to farmers in Iowa (Ryan and Gross, 1943), medical doctors in a large city (Coleman, Katz 14 and Menzel, 1966) or mental hospitals throughout the U.S. As will be noted later, careful consideration of the social system being examined and the units which define the system are essential parts of any diffusion research. Diffusion and a highly related concept, adop- tion, are best differentiated through the use of diffu- sion and adoption curves. The diffusion curve indicates the rate of adoption of an innovation by the adopting group in the social system. Graphically, Rogers (1971) and Havelock (1971) present this curve as follows: 100 84 Cumulative Percentage Late majority Adopting 50 Early majority 16 #Z/////, Early adopters 3 Innovators TIME Figure l. The Diffusion Curve (Havelock, 1971). 15 This indicates that it is possible to categorize groups of adopters in terms of their rate of adoption. Inno- vators comprise the first 3% to adopt the innovation. When 50% of the population has adopted, early majority adopters are included. Until nearly total adoption, however, it is difficult to categorize individuals in this manner. One possibility for categorizing individuals prior to total adoption is with an adopter curve. This represents the activity of an individual (or individual unit) as he adopts the innovation. We return to Havelock (1971) for one possible interpretation of the adoption curve: innovation becomes routine Cumulative . . efforts to adapt 1nnovat10n Involvement information-seeking slight involvement-awareness TIME Figure 2. The Adoption Curve (Havelock, 1971). 16 Rogers (1971) modifies this adoption curve and his adap- tation of it will serve as a basis for investigating the diffusion and adoption of the community lodge. Havelock (1971) and others discuss many charac- teristics of the receivers of innovation which influence their facility for adopting innovations. Past diffusion research examines three general elements: individual, interpersonal, and organizational. Organizational ele- ments will be investigated in the present study. These elements are traditionally divided into three categories: input (entering information), throughput (internal pro- cessing) and output (exiting information) (Havelock, 1971). Each element is important in the diffusion of an innovation. In order to examine the relationship between the diffusion of the lodge and certain organizational (hospital) characteristics, the present study was designed to test the following hypotheses. Hypotheses Hypothesis 1. Havelock (1971) and Katz and Kahn (1966) discuss specialized knowledge-seeking subunits whose aim is to seek out and collect new knowledge. Knoerr (1963) cites the example of library funds as important in facilitating the input of new knowledge. Others mention information retrieval systems (Veyette, 1962), planning units or 17 research and development units. Katz and Kahn (1966) stress the importance of such subunits having their own staff, having leaders with high status and having them report directly to the top of the organization. Therefore, we hypothesize that there is a significant ppsitive relationship between the existence of special- ized roles for input of new prpgrams and degree of pro- gress towards ad0ption of the lodg_. Hypothesis 2. Havelock (1971) and Katz and Kahn (1966) dis- cuss goals and rewards as they relate to information seeking. Rogers uses similar concepts in discussing "relative advantage." (Rogers, 1962, 1971.) An organ- ization which sees innovation and input of new knowledge as a part of its goal orientation, will operationalize such an orientation with rewards for innovative ideas, evaluations based upon innovativeness and public state- ments of knowledge-seeking goals. It follows that such an orientation would lead to greater progress towards adoption of the lodge innovation. Therefore, we hypoth- esize that there is a significant positive relationship between a hospital's_goals of seekipginformation about new programs and dggree of progress towards adoption of the lodge. 18 Hypothesis 3. Griffiths (in Miles, 1964) proposes a theory of innovation based upon change in leadership in an organiza- tion. His propositions include: 1) The number of innova- tions is inversely proportional to the tenure of the chief administrator; 2) Change in an organization is more probable if the successor to the chief administrator is from outside the organization, than if he is from inside the organization. Carlson (1962, 1965) has investigated change with respect to school systems and the superintendents of schools. He concludes that school superintendents "look upon a long tenure in office as detrimental to the school systems." (Carlson, 1965.) Change in superintendents is necessary, and the most beneficial is when the new leader comes from outside the system. "The outsider's performance changes the office and relations of others to the office . . . (he has) the tendency to change the district." (Carlson, 1965.) In fact, if the new superintendent has moved from community to community, he has ”placed himself in a group dedicated to change, reform and improvement." (Carlson, 1965.) Belknap (1956) came to a similar conclusion regarding mental hospital superintendents. Finally Marrow gt El: (1967) describe the complete change in leadership of an organization and cite the re— sulting infusion of new ideas. In addition to new ideas 19 at the top of the organization, the rest of the employees were psychologically better prepared for changes which followed. Therefore, we hypothesize that there is a signif- icant positive relationship between change in leadership and amount of_progress towards adoption of the lodgg. Hypothesis 4. In our brief discussion of the problem-solver perspective, it was mentioned that the organization must feel discomfort before new knowledge will be utilized (Lippitt, Watson and Westly, 1958; Lewin, 1951; Schein and Bennis, 1965). Schon (1967) states that "something like a state of crisis must arise . . . Once it perceives the threat, the organization must immediately interpret it as requiring a shift toward innovation." (p. 127) Watson (1966) indicates that innovation is resisted if things are going smoothly. Finally, Michael (1965, in Iiavelock, 1971) preposes that social disaster often fa- cilitates innovation. Therefore, we hypothesize that there is asigpi_- icant_positive relationship between occurrence of a crisis 1J1 the hospital and amount of progress towards adoption of the lodge. 20 Hypothesis 5. Fairweather, Sanders and Tornatzky (1972) found that several indices of broad decision-making participa- tion were consistent factors important in change related to the lodge during the approach-persuasion and implemen- tation phases. Change was related to a greater number of staff involved in the decision to implement the lodge, T If and a greater satisfaction with that decision. In addi- t tion, Griffiths, (in Miles, 1964) concludes that the more hierarchical the structure of an organization, the less the possibility of change. We hypothesize that there is a sigpificantgpositive relationship between a broad pat- tern of decision-making in the hospital and amount of progress towards adoption of the lodgg. Hypothesis 6, Organizations which feel that their programs are loetter than other hospitals, and are proud of the progress ‘their institution has made will be less likely to accept (Jutside innovations than those which are dissatisfied. ILavelock (1971) cites two studies which support this con- cilusion, Allen (1966) and President's Conference on Technical-Distribution Research (1957). We hypothesize ghgiz there is a significant negative relationship between :l9C411 pride" expressed by a hospital and_progress towards éioption of the lodgg. 21 Hypothesis 7. If a hospital in general feels strongly about the value of change, and if the superintendent, in par- ticular, feels that innovation is good and part of his job, then the lodge should diffuse more in that institu- tion. As an attempt to look at the general "zeitgeist" of the hospital as it relates to a specific change, we hypothesize that there is a significant positive rela- tionship between ”change orientation" in the hospital and amount of_progress towards adoption of the lodg_. Hypothesis 8. Katz and Kahn (1966) have developed the theo- retical notion of "systemic perspective." "Systemic research" involves the search for new information and a concern for the functioning of the organization in relation to its environment. It is governed by the 1optimizing principle, i.e., seeking not the minimally zaccepted answer, but rather the optimal answer. The Zlodge has been presented as the optimal answer to mental laealth.care, and it is concerned with the hospital's re- lationship to environment outside the institution. In addition, leaders with a systemic perspective are seen as “Killing to originate structure and to "change in re- Spouse to external demands for change." (Katz and Kahn, 1966..) We hypothesize that there is a significant 22 positive relationship between a "systemic" perspective envisioned bypthe administration of a hospital and pro- gress towards adoption of the lodg_. CHAPTER II METHODS SamplingUnits Two hundred and fifty-five State and V.A. hos- pitals for the mentally ill in the U.S. served as the sample for the present study. This is virtually the entire population of such hospitals. The sample in- cluded the following hospitals: 1. One hundred and seven hospitals who answered No to the persuasion condition and No_to the imple- mentation condition of the previous hospital imple- mentation study (Fairweather, Sanders and Tornatzky, 1972); henceforth to be called the NO/NO hospitals. 2. One hundred and twenty-three hospitals who answered ng_to the persuasion condition and No to the implementation condition; henceforth to be called the YES/NO hospitals. 3. Twenty-five hospitals who volunteered to im- plement the lodge by answering {gs to the implemen- tation decision. This includes two hospitals who answered N2_to the persuasion condition and 23 ‘- vv 0 ' _ 1" u 24 twenty-three who answered Yes; henceforth to be called the NO/YES and YES/YES hospitals, or imple- mentation volunteers. Of the 255 hospitals, 244 responded to the questionnaire, providing a return rate of 96 percent. Since all inter- viewing was completed by phone, we established a hier- archy of those to be contacted at each hospital. For each sample as described above, the interviewer prefer- ences were as follows: 1. 107 NO/NO hospitals a. Implementation decision maker b. Present superintendent 2. 123 YES/NO hospitals a. Implementation decision maker b. Other administrative old contacts c. Present superintendent 3. 25 NO/YES, YES/YES hospitals a. Research team's last contact b. Other administrative old contacts c. Present superintendent Data Collection Procedures The measurement of the variables in this study was based on three sources of information: a phone questionnaire, data from the previous hospital imple- mentation study, and demographic information available 25 from the American Hospital Association journal, Hospitals. The specific items of the questionnaire designed to mea- sure each variable are presented in Appendices B and C. This survey was conducted by telephone for sev- eral reasons. Expense ruled out site visits to all 255 hospitals. Written questionnaires have unreliable re- turn.rates, especially with a national sample. The phone contact allowed the researchers to gain a subjective "feel" for each of the institutions interviewed. Twelve questionnaires were mailed to respondents who requested it and who refused to answer by phone. Of those, two were completed and returned. The procedure used by the interviewers was as follows. Except during the Bell Telephone employee strike which occurred midway through data gathering, phone calls were made person-to-person in order to in- crease the likelihood of speaking with the preferred con- tact. During the strike, which lasted about 10 days, calls were made station-to-station. The phone introduc- tion used appears in Appendix A. In all cases the first contact as listed above was requested. After this there were three possibilities: 1. He was in and answered the call In this case the questionnaire is administered to him. If he referred us to someone else to get more or better information, contact was extended to the new 26 person. This was, in fact, a rare occurrence. Most contacts felt comfortable and confident in answering our questions. 2. He was still at the hospital, but unavail- able at the first call A message was left for him to call us collect (at specified hours). If contact was not completed by the second such attempt, the second person on the list was requested. This procedure was con- tinued until contact was made, and the information obtained. 3. He was no longer at the hospital Interviewers continued down the list as in l. and 2. In all cases, the status of the respondent was recorded and later investigated in the cluster analysis. Interviewers were three graduate assistants from the Department of Psychology at Michigan State University, including this investigator. In order to assure that each ‘presented the questionnaire from the same perspective, each interviewer was completely briefed on the lodge proj- ect and was allowed to read the files of his assigned hos- pitals. This gave each interviewer information necessary to administer the questionnaire including the names of Past contacts, persuasion’condition, superintendents' qunes and a feeling for the relationship between the 27 research team and the hospital. In many cases this pre- pared the interviewer for otherwise unexpected and con- fusing responses. Hospitals were randomly assigned to each interviewer. Prior to actual research interviews, each inter- viewer made calls to several hospitals not in the sample in order to acquaint him with the phone questionnaire and naturalistic interviewing. These practice inter- views were discussed in detail by the three researchers, and any misunderstandings were cleared up. In addition, the interviewers met for a brief session each day to en- sure questions were being asked identically and no prob- lems were arising. In general, interviewing went smoothly, though certain interviews required some persistence and patience on the part of the research team. Respondents were generally cooperative, if not friendly. Even though interviews averaged about 40 min- utes, there were few hostile complaints from the hospital jpersonnel. Some offered to send program descriptions to 115. Many wanted the results of the study sent to them. (lne interviewer was actually requested to interview for a. job at the respondent hospital. In spite of the above precautions, careful exam- iIlation of the responses elicited by each interviewer lead this investigator to suspect an irregularity with one iIlterviewer. These suspicions were confirmed when chi-square 28 tests indicated that this one interviewer received re- sponses significantly different from the other two inter- viewers on 11 of 20 variables tested. Therefore, an ad- ditional variable was inserted in the cluster analysis to test whether or not interviewer differences were related to the results obtained. Measurement of Variables ' As stated earlier, three sources of information 1} were used in this study: a phone questionnaire, data " from the previous hospital implementation study and demo- graphic information available from the American Hospital Association journal, Hospitals. The use of each of these is discussed below. Hospital Study Data Data deemed relevant from previous research (Fairweather, Sanders, and Tornatzky, 1972) included the :following items: 1. Social change score (1-3) 2. Persuasion condition (brochure, workshop, demonstration ward) 3. Persuasion volunteer (Yes, no) 4. Implementation volunteer (Yes, no) Each of these had been scored previously and fit well into the analysis of the present study's data. 29 Demographic Information Demographic data was available from the AHA journal, Hospitals. In addition to its face value, this demographic data gave us an objective look at several of the "crises" mentioned by respondents. Crises usually involved budget or staff, and therefore an objective mea- sure of such crises was available from the following data: Total expenses 1970 Difference in expenses 1969-1970 Total number of staff 1970 Difference in staff 1969-1970 Staff/Patient ratio 1970 Difference in census 1969-1970 GUI-RUIN!“ Phone Questionnaire The phone questionnaire was divided into three distinct sections: lodge diffusion questions, hypothesis- testing questions and innovative program descriptions. 'The measurement of each of these variables is discussed below. Lodge Diffusion Questions The specific items designed to measure diffusion 01? the lodge are presented in Appendix B. Rogers (1962, 15971) provided the basic framework for these questions. Tllis framework was used to index the extent of lodge adaption by individual hospitals. Since few of the 30 hospitals had actually adopted the lodge, a score of progress towards adoption was more appropriate than a score based upon the traditional diffusion curve (see Appendix D). Questions were therefore asked in terms of Rogers'(l97l) stages of adoption: 1. Knowledge 2. Persuasion 3. Decision 4. Confirmation The resulting score (Diffu) is a measure of movement towards adoption of a lodge by each hospital since the implementation decision date. Since this was the last contact with the 230 NO/NO and YES/NO hospitals, it is also a measure of diffusion for those hospitals. It is not an uncontaminated diffusion score for the 25 implementation volunteers since it includes the time period during which implementation attempts ryere being made. A second score (Diffu 2) was there- :fore created to investigate movement towards adoption cxf a lodge since our last contact with the volunteer IHDSpitals. Both of these measures are presented in de- tail. in Appendix D. Two interviewers reached interrater reliability of .92 on Diffu and .82 on Diffu 2 (Pearson PITDduct moment correlation coefficient.) 31 Hypothesis—Testing Questions A series of questions was presented to each re- spondent to test the hypotheses listed above. The spe- cific items are presented in Appendix C. Questions were designed to investigate the following concepts: 1. Innovative In an attempt to compare lodge adoption activity ivith a measure of general adoptiveness and innovativeness, each respondent was asked to describe treatment programs Type of subunit designed to search for new programs. Types of activities funded by the hospital. Goals expressed by the hospital. Change of administration in the hospital. Crisis situations in the hospital. Power and type of influence in developing new programs. Subjective assessment of the hospital's treatment facilities. Systemic research approaches by the hospital staff. Program Descriptions Sitarted at his hospital since our last contact with it. III general, this was since April, 1969. The 25 volun- tiaers were asked to describe any new programs started since they had volunteered to implement the lodge. FI‘om these descriptions, and from a checklist read to 32 each respondent, the following new programming scores were developed: 1. Total number of new programs. 2. Number of new in-hospital programs. 3. Number of new community programs. 4. Degree of community locus exhibited by new programs. 5. Degree of autonomy for patients exhibited by new programs. These measures are fully described in Appendices E and F. Interrater reliability was again calculated employing a Pearson product moment correlation coefficient. The re- sults were as follows: 1. Total new programs r = .86 2. In-hospital programs r = .93 3. Community programs r = .77 4. Community locus r = .87 5. Autonomy r = .98 Mean r = .88 ffior our purposes, "program" was any organized activity 01? group of activities whose focus is on patients from tile hospital. This would not include outpatient clinics ffrr community patients, the "unit" system per se nor "irraining" for staff. Therefore, some judgement was uS ed in. deciding what would be considered a program. 33 Data Analysis The basic outcome criterion was the diffusion score (Diffu) described above which indicates the de- gree to which each hospital adopted the lodge since the implementation decision date of the previous hospital study (Fairweather, Sanders and Tornatzky, 1972). Since the experimental attempts to approach, persuade and im- plement the lodge may logically have an effect upon the degree of adoption, a chi-square test was made using the diffusion score and previous willingness to adopt the lodge concept. In this manner a comparison was made of unplanned diffusion (NO/NO hospitals), partially planned diffusion (YES/NO hospitals) and planned diffusion (YES/ YES, NO/YES hospitals) on degree of adoption. Our hypotheses predict positive relationships between degree of adoption and eight concepts tested during the study. These eight concepts were defined .such that, along with some defining variables, the total Jiumber of variables investigated in this study is 90. 131 order to handle this amount of data, and to determine any associative relationships between degree of adoption aJLd.the other variables measured in this study, Tryon and Bailey's (1970) methods of Cluster Analysis were used. SPe-‘ecifically a V-Analysis was used to define the empirical ‘Illasters. A preset analysis was then employed with the diffusion score (Diffu or Diffu 2) as the key variable ‘31:?" ‘ 34 in the first cluster and the definers of the empirical V-Analysis preset for succeeding clusters. Variables are included in the CC5 and CSA programs of V-Analysis which are most highly collinear with other members. In general, variables are excluded whose factor loading is below .40, and whose communality is below .20. I ~‘ ”Ia—5‘" CHAPTER III RESULTS Amount of Diffusion i The amount of diffusion as measured by the de- FEE—9“ gree of adoption score (Diffu) is the first concern of this chapter. Many analyses reported in this section are based upon the assumption that diffusion did in fact occur, and that this diffusion has significant re- lationships with groups of variables measuring the con- cepts cited in our hypotheses. An analysis of the amount of diffusion is pre- sented in a chi-square table using the diffusion score (Diffu) and nominal categories which indicate the hos- ;pital's previous willingness to adopt the lodge concept. 'The influence of the latter categories was not hypothe- :sized, but such a distinction is a logical one for this analysis. We must remind the reader that Diffu is a Scuare created to indicate the degree of adoption reached b)’ a hospital regardless of previous experimental condi- 'tixons. It is a diffusion and dggrge of adoption score for 219 hospitals (NO/NO and YES/NO hospitals) and only 35 36 a degree of adoption score for the 25 volunteer hospitals (YES/YES, NO/YES hospitals). Diffu 2 is the diffusion score for this latter group of hospitals (see Appendix D). Table 2. Comparison of Amount of Diffusion Between Hos- pitals Indicating Differential Willingness to Adopt the Lodge in Previous Research Efforts. ‘ previous Amount of Diffusion (Diffu)» Willingness to - ° - Adopt Lodge Knowledge Persuasion Riggiig: (1) (2) (3) (4) (5+) None (NO/NO) 68 18 s 6 2 Permitted Persuasion Attempt (YES/NO) 52 27 l6 l6 6 Volunteered to Adopt Lodge (NO, YES/YES) 0 S 2 3 15 x2 123.9 df =,3 p <.001 The most striking result is the lack of diffusion- adoption. Only 23 hospitals of the 244 investigated ac- tually proceeded beyond discussion of the lodge concept. Of these 23, 15 were volunteers who regularly received input from the research team. The mean diffusion score 37 is 2.25. For the NO/NO hospitals it is 1.62; for the YES/NO hospitals it is 2.20; and for the volunteer group it is 5.08. For the latter group, eight hospitals adopted the lodge in some fashion (a "7" or "8" on Diffu) while only one hospital reached this level in the other 219 hos- pitals. The chi-square computed on these scores is signif- icant beyond the .001 level, indicating that the three groups of hospitals are significantly different with re- gard to degree of diffusion-adoption. In other words, de- gree of diffusion-adoption is related to previous implemen- tation attempts. Because of these significant differences cluster analyses which follow have been computed indepen- dently for the NO/NO, YES/NO and implementation volunteer hospitals. Diffusion and Degree of Adoption - A Closer Look at t e 25 Volunteer Hospitals We have indicated the distinction between diffu- sion and degree of adoption for the 25 volunteer hospitals (Diffu vs. Diffu 2), but have not as yet looked at the re- lationship between the two measures. Figure 3 indicates this relationship. This graph indicates two things. First, that a limited amount of diffusion (Diffu 2) did occur in this group. And secondly, that the diffusion occurred most markedly in those hospitals which adopted the lodge. 38 coapmehwmdou scamfiooo dowmmsmuom w n :owumow< mo oonwom o d .mampwmmom umouc3H0> mm one pom cowpaow< mo oonwon wow scamsmmwa mo :omfiummEou .m ouswfim ”scones m d v m 11 N owvoazoau H d :ofimmsomao Haaonpueea oz RN steamy coauum 30¢ 02 nowmsmwwm .nowmmsomwo zoo: osom cofiuum zoz 39 Cluster Analysis of Diffusion and Hypothesis~TestingTVariables An empirical V-Analysis (Tryon and Bailey, 1970) was done in each of the three groups of hospitals as des- cribed in the last section (NO/NO hospitals, YES/NO hos- pitals and YES/YES or NO/YES hospitals). The three sep- arate analyses were decided upon due to the differences reported in Table 2. The results of the empirical V- Analysis failed to form a cluster which included the diffusion score (Diffu or Diffu 2) for the three groups of hospitals. These variables were rejected from the cluster analysis due to the fact that their communalities were below .20. The diffusion score appeared as a sep- arate and distinct variable. This finding is itself an important one and will be discussed at length in Chapter IV. In order to form a diffusion cluster, we then preset the diffusion score (Diffu or Diffu 2) and other cluster definers derived from the empirical analysis (Preset key-cluster analysis, Tryon and Bailey, 1970). In this way we are able to analyze our data with respect to specific variables related to diffusion and also in- vestigate variables independent of that concept. The preset clusters which appear in the NO/NO and YES/NO hospitals had reliabilities of only .21 and .27 respec- tively. The factor loadings for the diffusion score 40 were relatively low for definers, .46 (NO/NO) and .52 (YES/NO). The 25 volunteer hospitals did, on the other hand, present a more reliable cluster. This is one in- dication that where diffusion dogs occur certain vari- ables grg_related to it; and that its unreliability in the first two analyses is probably due in large part to the fact that little diffusion actually occurred in those hospitals. The NO/NO Hospitals The clusters obtained from those hospitals which indicated no willingness to adopt the lodgg (Fairweather, gt_§i., 1972) are presented in Table 3. The correlations between oblique cluster domains (correlations between the rotated oblique factors) are presented in Table 4. Diffusion This cluster is entitled "diffusion" because of the pre-set nature of the diffusion score (Diffu). The weakness of the cluster is immediately apparent due to the diffusion score's low factor loading (.46), a reli- ability of only .21, and the disparate nature of the variables included. Recalling the lack of diffusion re- ported above this is not an unexpected finding. Figure 4 reveals a non-normal curve not likely to enhance the 41 Table 3. The Ten Clusters in the NO/NO Hospitals. AA Cluster Loading Cluster 1. Diffusion l.'Diffusion score (Diffu) .46 2. Low status head of group looking for new programs .63 3. More funds for workshops .50 4. Hospital has little concern for community needs .44 Cluster 2. "Middle" discipline total influence on new ro rams 1. Social work: a. Greater breadth of influence .85 b. Greater total influence .88 2. Psychology: a. Greater breadth of influence .88 b. Greater amount of influence .60 c. Greater total influence .86 3. Nursing: a. Greater breadth of influence .66 4. Vocational rehabilitation: a. Greater breadth of influence .68 b. Greater total influence .66 5. Higher mean breadth of influence for hospital .86 Cluster 3. Hospital census data ’1. Gfeater increase in census (less decrease in census) 1.00 2. Less absolute difference in census, 1969-1970 .84 3. Less decrease in occupancy, 1969- 1970 .76 4. Small number of total staff .35 5. Low staff/patient ratio .32 Cluster 4. Superintendent influence on new proggams l. Superintendént: a. Greater total influence 1.00 b. Greater breadth of influence .87 c. Greater amount of influence .70 2. The most influential discipline is a high status position. .54 Cluster 5. Hospital expenses 1. Greater increase in exPenses, 1969- 1970 .96 2. Greater absolute difference in ex- penses, 1969-1970 .93 42 Table 3. Continued Cluster Loading Cluster 6. Adpption of Innovative programs 1. New Programs a. Greater total number 1.00 b. More in-hospital programs .81 c. More community programs .45 d. Higher community locus score .53 e. Higher autonomy score .38 Cluster 7. Psychiatrist influence on new programs 1. Psychiatrist: a. Greater breadth of influence .95 b. Greater total influence .84 2. Greater resistance in hospital .51 Cluster 8. Crises _1. A crisis has occurred in the hos- pital since our last contact .93 2. More crises have occurred since our last contact .92 Cluster 9. Amount of Influence on new programs ’1. Greater mean amount of influence across all disciplines .89 2. Greater amount of influence: a. Social work .79 b. Nursing .73 c. Vocational rehabilitation .61 d. Psychiatry .41 3. Greater total influence: a. Nursing .65 4. Lower variance of the amount of influence .53 Cluster 10. Committee to find new programs I. More people on the comm1ttee .92 2. More disciplines represented on comm. .73 3. Less time spent in this function by head of the group .42 4. Superintendent came to position from outside the hospital .57 5. Superintendent is interested in public relations .48 ‘1! I. mags-w ‘4 43 - NH.- NO. NH.- OH. OO.- HO.- HH.- ON.- OH.- OopuHeaou seaweed :82 .OH NH.- - OH.- NN. NH. HH.- OH. OH. OO. NO.- mucosHHOH Ho Hesoe< .O NO. OH.- - OO. OH. NH.- OH.- NH.. OH.- OH. momHNo .O NH.- NN. OO. - HO. NO. ON. OO.- mm. OO. mucosHHOH OmHupangsmm .N OH. NH.- OH. HO.- - OO.- OH.- NN.- OO. NO. mamamsHOO>oeeH-mmocosHOOou< .O OO.- HH.- NH.- NO. mO. - HH. HO.- OO. HO.- momeogxo HOOHOmOm .m HO.- OH. OH.- OH. OH. HH. - ON. ON. ON.- OOOOOHHOH HOOOOOHOHumaam .4 HH.- OH. NH.- OO.- NN. HO.- ON. - NO.- NN. OHOO mamaou HOOHOmOm .N ON.- OO. mH.- mm. OO. NO. ON. NO.- - OO. HOHON oeHHOHummwcwwwwmw .N OH.- NO.- OH. NO. NO. HO.- ON.- NN. OO. - OonsmmHO .H OH a m N o m e m N H muopmsHu .mHmuHmmom oz\oz ago How mdwwsoa Houmdau odcwaoo doozuom mdowpmaouuou .v canoe 44 necessary collinearity upon which both factor loading and reliability are based (Tryon and Bailey, 1970). 70 " Number 50 n of Hospitals 30 .. 10 L i 2 3 4 5+ Knowledge Persuasion Decision-Adoption Amount of Diffusion Figure 4. Amount of Diffusion in the NO/NO Hospitals. The moderate relationship between this cluster and other cluster domains, Adoptiveness-Innovativeness (.42) and Hospital Expenses (-.41), and Superintendent Influence (-.26) must be looked at with these restric- tions in mind. Nevertheless, it appears that the little diffusion which did occur, was related to hospital con- Cepts other than just diffusion. This finding will be discussed more fully in Chapter IV. Middle Discipline Influence This cluster is essentially a group of variables measuring the breadth and total influence of social work, psychology, nursing and vocational rehabilitation in the hospital. The high relationship between this cluster, 45 Psychiatrist Influence (.53) and Amount of Influence (.44) reveals a group of variables which measure in- fluence for all disciplines except the Superintendent. Hospital Census Data This cluster is a group of demographic vari- ables descriptive of hospital census information. Mildly related to the census data are two staff vari- ables. These indicate a low staff/patient ratio, and a lower number of staff. This cluster is only mildly related to two other clusters, Superintendent Influence (.29), and Adoptiveness-Innovativeness (-.22). Superintendent Influence This cluster includes the variables which mea- sure the superintendent's amount, breadth and total in- fluence in the hospital. It is interesting that this cluster is only mildly related to Middle Discipline Influence (.23) and moderately related to Psychiatrist Influence (.34). There appear to be three levels of influence reported in our sample, superintendent, psy- chiatry and "others." This cluster is also mildly re- lated to Hospital Census (.29). 46 Hospital Expenses This cluster consists of two variables which indicate an increase in expenses from 1969-1970. Its relationship with other clusters is negligible. Adoptiveness-Innovativeness This cluster includes all variables which in- vestigate new programs recently adopted in the hospitals. It includes both the quantitative and qualitative mea- sures, and indicates that the greater the number of pro- grams, the more likely one of them will be innovative. It is relatively independent of other hospital clusters, and only mildly negatively related to Hospital Census (-.22), though it is related to the diffusion cluster (.42).- This may indicate that the greater diffusers are also the most generally adopting. However, the low reliability of the diffusion cluster limits the inter- pretability of this finding. Psychiatrist Influence This cluster includes breadth and total influence variables of psychiatry in the hospital. It is interest- ing that it also includes the measure of greater resis- tance to new programs in the hospital. It would appear that greater influence of psychiatry is associated with greater resistance. The high relationship between this V-‘E-iu . 47 cluster and Middle Discipline Influence (.53) has been mentioned above. In addition, this cluster shows mod- erate relationship with the Superintendent Influence cluster (.34). Crises This cluster of two variables indicates the occurrence of a crisis in the hospital as reported by the respondent. Its relationship with other clusters are negligible. Amount of Influence This cluster contains variables descriptive of the amount of influence accorded social work, nursing, vocational rehabilitation and psychiatry as well as a greater mean of influence for new programs across all disciplines. In addition, it includes the variable in- dicating lower-variance of influence scores and thus a kind of flatness of power. This cluster's relationship with Middle Discipline Influence (.44) and Psychiatrist Influence (.23) have been discussed above. New Program Committee This cluster includes variables which describe the committee for new programs and also two superinten- dent variables. The variables indicate a larger 48 committee, a superintendent interested in public rela- tions and a superintendent who came from outside the hospital. This cluster is essentially unrelated to other clusters. The YES/NO Hospitals The clusters obtained from the hospitals which indicated a willingness to receive a brochurep attend a workshop or develop a demonstration ward, but would not volunteer to implement the lodgg_are presented in Table 5. In Table 6 the correlations between oblique cluster domains (correlations between the rotated ob- lique factors) are presented. Note at the outset the striking resemblance to the clusters of the NO/NO hos- pitals. Diffusion This cluster is subject to the same reserva- tions as expressed for the NO/NO hospital diffusion cluster. The factor loading of the diffusion score is only .52, reliability is .27 and once again the mean diffusion is low, 2.20 indicating an overemphasis at the low end of the diffusion scale. In this case, experimental conditions become variables in the cluster, including persuasion condition, interviewer and.respon- dent variables. These latter two variables merely 49 Table 5. The Nine Clusters in the YES/NO Hospitals. Cluster Loading Cluster 1. Diffusion l. Diffusion score .52 2. Experimental conditions a. More active persuasion condition .70 b. Lower status respondent .41 c. Interviewer was B.P. .58 3. Other a. More money for rewards .83 b. Higher status head of group looking for new programs .44 Cluster 2. Total influence of all disciplines below superintendent I. Sociai work a. Greater breadth of influence .95 b. Greater total influence .88 2. Psychology a. Greater breadth of influence .80 b. Greater total influence .73 3. Nursing a. Greater breadth of influence .75 b. Greater total influence .70 4. Vocational Rehabilitation a. Greater breadth of influence .69 b. Greater total influence .61 5. Psychiatry a. Greater breadth of influence .68 b. Greater total influence .53 6. Greater resistance to new programs .51 7. Greater mean amount of influence .92 annster 3. Hospital size 1. Larger staff .96 2. Larger budget .96 3. Greater difference in census, 1969- 1970 .66 4. Greater difference in staff, 1969- 1970 .42 gills ter 4 . Adoptiveness - Innovativenes s 1. New programs a. Greater total number 1.00 b. More in-hospital programs .81 c. High community locus score .44 2. Committee to find new programs a. Meets rarely .52 b. Greater diffusion of its informa- tion .40 50 Table 5. Continued Cluster Loading Cluster 5. Hospital expenses 1. Greater increase in expenses 1.00 2. Greater absolute difference in expenses .91 Cluster 6. Superintendent influence on new progggms l. Superintendent a. Greater amount of influence .88 b. Greater breadth of influence .66 c. Greater total influence .69 2. The most influential discipline is a high status position .74 Cluster 7. Amount of influence on new programs 1. Greater mean amount of influence across all disciplines .97 2. Greater amount of influence a. Social work .79 b. Nursing .69 c. Psychology .61 d. Vocational rehabilitation .60 e. Psychiatry .35 3. Low variance of the amount of in- fluence .62 Cluster 8. Crises 1. A crisis has occurred in the hos- pital since our last contact .90 2. More crises have occurred since the last contact .92 Cluster 9. Hospital census data 1. Greater increase in census (less decrease in census) .90 2. Less decrease in occupancy, 1969- 1970 .80 3. Superintendent is interested in community programs .86 4. Informal source of new programs has low status .42 51 - Ho. mH. NN. no. OO.- 0O. Nm. Oo. «new mamcoo HmuHmmom .m Ho. - 00.- mo.- mH.- Oo. No. OO.- 0H. momHHu .w mH. OO.- - 0H.- NH. mH.- HH. NO. NN. oodosquH mo pesoa< .5 ON. mo.- OH.- - mo. Oo. mo. HH. Ho. mucosHmcH ucowcoucHuomzm .0 NO. OH.- NH. OO. - OO.- OO.- OH. HO.- mmmeomxo HOOHOmom .m Oo.- Oo. mH.- Oo. mo.- - OH. mo. NH. mmooo>Hom>occH -mmo:o>Humov< .O 0O. No. HH. no. mo.- OH. - HN. mo. oNHm HmuHmmo: .m Nm. no.- OO. HH. mH. mo. HN. - OH. mucosHmnH Hmwoe .N Oo. oH. NN. Ho. Ho.- NH. mo. OH. - QOHmdmmHn .H m N N o m O m N H muoumsHu .mHauHmmom oz\mm> 0:» wow demsoa HopmsHu oddHHoo coozpom mGOHumHouuou .o oHomH 52 indicate that one interviewer was associated with hos- pitals which had higher diffusion scores and that the respondent in the greater diffusing hospitals was of lower status. However, the other experimental variable indicates an important finding. The more active the persuasion condition, the more diffusion which occurred in the hospitals which did, in fact, allow the persua- sion attempt. Apparently, the more active the process of involvement, the more intense will be future activity (diffusion). It is also of interest, that unlike the NO/NO diffusion cluster, this cluster is virtually un- related to other cluster domains. This will be discussed further in Chapter IV. Total Influence This cluster is identical to the Middle Dis- cipline Influence cluster of the NO/NO hospitals with the addition of the Psychiatrist Influence cluster (Table 3). It includes the breadth and total influence of all disciplines below superintendent, and the resis- tance measure. It is moderately related to both Amount of Influence (.47) and Hospital Census (.32). We see here a clear distinction between the superintendent and' other disciplines since the correlation between this cluster and the Superintendent Influence cluster is only .11. 53 Hospital Size This cluster did not appear in the NO/NO hos- pital clusters, though it is moderately related to the Hospital Census cluster (.46) which appears in both. This cluster is described by-variables indicating a large hospital in terms of staff, budget and differences in census and staff. It is interesting to note that hos- pital size is virtually unrelated to measures of Adoptiveness-Innovativeness (.14). Adoptiveness-Innovativeness This cluster is similar to the Adoptiveness- Innovativeness cluster presented previously (Table 3). It includes total new programs, more in-hospital pro- grams and the community locus score. It does not in- clude more community programs or the autonomy score. The addition of two variables concerned with the new program committee is interesting, though limited. This cluster's correlations with other clusters are negligible. Hospital Expenses This cluster is identical to the Hospital Ex- penses cluster reported earlier. Its relationship with other clusters is negligible. 54 Superintendent Influence This cluster, once again, is identical to the Superintendent Influence cluster presented in the NO/NO hospital analysis. The moderate relationship between this cluster and the Psychiatrist Influence cluster and Middle Discipline Influence cluster does not appear for these hospitals. Apparently the superintendent is seen as less a part of the other disciplines than in the NO/ NO hospitals. Only with Hospital Census (.27) does the correlation with another cluster reach even a mild re- lationship. Amount of Influence This cluster is another which is virtually identical to a cluster presented in Table 3. This in- cludes the amount of influence variables of all disci- plines except superintendent. It is moderately related (.47) to the Total Influence cluster. Crises ’ This cluster is identical to the Crises cluster in Table 3. It is, once again, virtually unrelated to all other clusters. 55 Hospital Census Data This cluster is similar to the Hospital Census cluster in Table 3, though it is not as consistent. It includes only two census variables, and adds one superin- tendent and one new program source variable. Its rela- tionship to the Total Influence (.32), Hospital Size (.46) and Superintendent Influence (.27) clusters have.been men— tioned previously. The YESLYES and NO/YES Hospitals (Volunteers) The clusters obtained from the 25 hospitals which indicated a willingness to implement the lodgg are presented in Table 7. This analysis was only done with the twenty-five hospitals which volunteered to adopt the lodge. This decreased N size has some implications for the interpretability of the obtained clusters, though it appears to offer some valuable leads. The diffusion score employed in this analysis is Diffu 2, while Diffu still indicates the degree of adoption of the lodge. In Table 8, we have presented the correlations between ob- lique cluster domains (correlations between the rotated oblique factors). Diffusion In this analysis a strong diffusion cluster appears. The diffusion variable (Diffu 2) has a factor 56 Table 7. The Three Clusters in the Implementation Volunteer Hospitals. Cluster Loading Cluster 1. Diffusion l. Diffusion-adoption a. More diffusion since last implementation attempt (Diffu 2) .88 b. Greater degree of adoption (Diffu) .80 2. Hospital power structure a. Less amount of influence by the superintendent .65 b. Less amount of influence by psychiatry .61 c. Lower mean amount of influence .54 3. Hospital is not generally for new programs .45 Cluster 2. Informationg gatherifgk- Stability TI .iPositive 1nforma on gathering a. More funds for workshops .98 b. More funds for travel .92 c. Funds exist to reward staff for new ideas .49 d. Hospital actively looks for new programs .75 e. There is a committee to look for new programs .45 f. More disciplines on the new program committee .51 g. More people on the new program committee .48 2. Crisis a. Few crises occurred .69 b. Little change due to crisis .56 c. A crisis did occur .43 3. Present programming a. All programs approved were im- plemented .65 b. Programs are seen as fine the way they are .49 4. Staff, expenses a. Small change in staff, 1969-1970 .71 b. More staff, 1969-1970 .67 c. Small numbers of total staff .54 d. Small budget .45 e. Low staff/patient ratio .41 57 Table 7. Continued. Cluster Loading 5. "Middle" discipline influence a. Greater amount of influence- psychology .61 b. Greater amount of influence- nursing .50 c. Greater breadth of influence-1 social work .61 d. Greater breadth of influence- psychology .49 e. Greater breadth of influence- vocational rehabilitation .48 f. Greater total influence-social work .58 g. Greater total influence-psychology .52 h. Higher mean breadth of influence .56 6. Experimental conditions a. More active persuasion condition .49 b. Interviewer was B.P. .52 Cluster 3. Adoptiveness-Innovativeness 1. Innovative programming a. Greater total number of new pro- grams 1.00 b. Greater number of in-hospital programs .85 c. Greater number of community pro- grams .55 d. Higher autonomy score' .53 e. Higher community- -locus score .52 f. Less resistance to new programs .43 2. Superintendent role a. Superintendent seen as innovative 55 b. Superintendent came to hospital from inside the hospital .48 3. Search for new programming a. New program committee does little with their information .82 b. New program committee meets rarely .71 c. Informal source for new programs is a low status person .71 d. Less money for the library .42 S8 loading of .88. There was some diffusion beyond mere discuSsion in contrast to our finding in the NO/NO and YES/NO hospitals. While only 4 percent (6 of the 219 hospitals) of the hospitals in those two conditions achieved a level greater than discussion, 20 percent (5 of 25) of the implementation volunteer hospitals made actual new movement towards adoption. Table 8. Correlations Between Oblique Cluster Domains for the Implementation Volunteer Hospitals. l 2 3 1. Diffusion - .04 -.11 2. Information .04 - .11 gathering- Stability 3. Adoptiveness- -.11 .ll - Innovativeness As indicated in Figure 3, greater degree of adoption (Diffu) is related to more subsequent diffusion. In addition, the power structure variables included in this cluster are interesting; less amount of influence for both superintendent and psychiatry, and a lower mean amount of influence for the hospital. It appears that lessened traditional hierarchical power is related 59 to diffusion and adoption. Yet the final variable in- cluded indicates that verbally the hospital does not en- courage new programs. This cluster is virtually unre- lated to either of the other two clusters. Information gathering - Stability This cluster includes several variables which seem to express a feeling of stability in the hospital. These include variables descriptive of few crises, sat- isfaction with present programming, small changes in staff numbers, and generally equal influence across psy- chology, nursing, social work and vocational rehabilita- tion. A second aspect of this cluster is the vari- ables which indicate a positive approach to information gathering. These include more funds for workshops, travel, and rewards for innovative ideas as well as vari- ables indicating a large committee to look for new pro- grams. Finally, experimental condition variables are also included, though they make little rational sense. This cluster is unrelated to the other two clusters. Adoptiveness~Innovativeness This cluster differs from the Adoptiveness- Innovativeness clusters of the previous two discussions in that it includes both programming variables and 60 variables describing the superintendent's role and gen- eral information-seeking. In addition to the innovative programming variables, we find variables describing the superintendent as innovative and as coming from within the hospital structure. It is interesting that the in- formation-gathering variables are negative ones; the new program committee meets rarely and does little with their information. The positive variables of information- gathering appeared in Cluster 2; and appear to be a part of a stable, not innovative system.. As indicated pre- viously, this cluster is unrelated to the other two clus- ters . Hypothesis-TestingVariables Table 9 is a summary of the cluster loadings for each hypothesis-testing variable on diffusion. All loadings over .40 are marked with an asterisk (*). What is immediately apparent is the lack of load- ings over .40, and the resultant disconfirmation of vir- tually all hypotheses in all three conditions. In fact, the only hypothesis which may be considered even par- tially confirmed is Hypothesis 5, for implementation vol- unteer hospitals. There is some indication of less hier- archical structure in decision-making regarding new pro- grams. The low amount of influence for superintendent and psychiatry, the low mean influence, and the general Table 9. on Diffusion Score. Cluster Loadings of Hypothesis-Testing Variables Cluster Loading Hypotheses Hospitals NO/NO YES/NO YES/YES, NO/YES Hypothesis 1. 1. Is there a group or in- dividual to look for \_ new programs? -.06 .20 -.31 2. No. of people in group -.03 -.05 .26 3. No. of disciplines in group. -.12 -.14 .13 4. What is done with in- formation? .61* .12 .18 5. How often does group meet? -.04 .03 .25(N=8) 6. Is there an informal contact for new pro- grams? .21 -.19 -.05 7. Status of informal contact -.27 -.15 -.S3*(N=6) 8. Status of head of group -.63* .44” .57* Hypothesis 2. 1. 15 reporting and receiving new infor- mation a goal of your hospital? .20 .36 -.27 2. Do staff members pre- sent programs? .26 .21 .08 3. Are there funds for: a. Workshops .50” .06 .02 Amount .14 -.31 .19 62 Table 9. Continued. Cluster Loading Hypotheses Hospitals NO/NO YES/NO YES/YES, NO/YES b. Library .18 .12 .23 Amount .28 .06 .10 c. Travel .27 -.22 -.20 Amount -.19 -.02 -.12 d. Rewards .13 .13 -.32 Amount -.44* .83“ .43*(N=9) 4. Relative importance of innovative ideas .05 .23 .30 Hypothesis 3. 1. Have any new adminis- trators joined the hospital? -.14 .02 .17 2. No. of positions changed -.07 .15 .22 3. No. of years of super- intendent in office -.13 .13 .25 4. Did superintendent come from outside the hospital? -.21 .17 .04 Hypothesis 4. 1. Has there been a crisis? .24 .21 .09 2. No. of crises? .04 .08 -.06 3. Was the crisis resolved? .30 .05 .04 4. How much change due to crisis?; .10 -.33 -.47* (Demographics) 5. Increase in occupancy rate , .14 .09 .03 63 Table 9. Continued- Cluster Loading Hypotheses Hospitals NO/NO YES/NO YES/YES, NO/YES 6. Increase in census .17 -.01 .04 7. Increase in expenses -.42* .03 -.08 8. Increase in staff -.08 -.32 -.05 Hypothesis 5. 1. Amount of influence a. Supt. -.29 .ll -.65* b. Psychiatrist .00 -.07 -.61* c. Psychologist .21 -.07 -.26 d. Social Worker -.11 .30 -.22 e. Nursing .11 .21 .06 f. Vocational Rehabili- tation -.09 -.07 .27 2. Breadth of influence a. Supt. .07 -.04 -.34 b. Psychiatrist .09 .17 .12 c. Psychologist .06 .08 .10 d. Social Worker .03 .14 .20 e. Nursing .10 .25 .35 f. Vocational Rehabili- tation .23 -.05 .31 3. Total influence a. Supt. .06 -.03 -.42* b. Psychiatrist -.01 .ll -.32 c. Psychologist .19 -.02 -.10 d. Social Worker .00 .23 .08 e. Nursing .15 .27 .31 f. Vocational Rehabili— tation .26 -.18 .25 4. Mean.breadth of influence .13 .14 .23 5. Mean influence -.04 .10 -.54* 6. Variance of influence .31 -.07 .12 7. Amount of resistance .23 .10 -.02 1.1a lII I.I|.>IIII IIII iL.I...nNLm.Ij~ 64 Table 9. Continued. Cluster Loading HYP°th9535 Hospitals NO/NO YES/NO YES/YES, NO/YES 8. Supt.-Role a. leader .00 .19 .00 b. innovative pro- gramming -.17 .37 .24 c. community pro- gramming -.13 .02 -.17 d. public relations .07 .13 -.25 Hypothesis 6. 1. Positive subjective assessment of hospital .26 -.12 -.05 2. How much better should programs be? -.35 .06 -.05 Hypothesis 7. 1. Is the hospital for ' starting new programs? -.24 .18 -.45* 2. Any programs approved, but not implemented? .26 .04 -.37 3. Any programs discon- tinued? .12 -.03 -.25 Hypothesis 8. . Systemic l. Concerns a. IocaI community -.16 .00 -.35 b. optimal answer -.30 .12 -.32 c. evaluation .20 -.09 -.17 d. community needs -.44* .07 -.26 e. active search for new programs -.04 .25 -.33 1r .1. 65 pattern of loadings, reveals some degree of influence for those disciplines below the traditional strong ones, par- ticularly nursing. CHAPTER IV DISCUSSION Diffusion Amount of Diffusion The lack of diffusion of the lodge concept is the most distinguishing characteristic of the 219 hos- pitals which had initially refused to implement the lodge. Other writers have warned that innovations do not spread automatically (Glaser, 1967). It appears that innovations such as the lodge only spread with.a maximum effort from the research team's action consul- tants. Anything less than full scale implementation attempts (as with the 25 volunteer hospitals, see below) seems to result in little or no significant diffusion of a complex concept. The amount of diffusion since the last contact with the non-volunteer hospitals reveals no more than the difference between awareness of the lodge and ig- norance of it. Over half of these respondents revealed no awareness of the lodge program. Of those aware of the concept, only eight indicated movement beyond 66 67 discussion. Little lodge diffusion occurred automati- cally; and it only rarely occurred even in those hos- pitals which had been exposed to persuasion attempts (YES/NO hospitals). The 25 hospitals which volunteered to implement the lodge are markedly different from the 219 hospitals justdiscussed. This has been examined earlier in the present study. The major experimental difference is that each of these hospitals was subjected to implemen- tation efforts beyond the final contact with the other 219 hospitals. There were planned attempts made to im- plement the lodge. Therefore, "diffusion" for these hospitals, while parallel to diffusion for the non- volunteers, requires a revised definition (see Appendix D).‘ The diffusion score (Diffu 2) here measures prog- ress towards adoption of a lodge 31222 the last imple- mentation effort; e.g., for those hospitals which had actually implemented the lodge, this was a measure of progress towards adoption of a second lodge. For this group of hospitals, some diffusion did occur, as shown in Figure 3. It is still minimal, but discernible. Figure 3 also reveals that the diffusion which did occur, took place in those hospitals which had adopted the lodge during implementation attempts. For the most part, those who stopped short of adopting the lodge during active implementation efforts, did not 68 complete adoption once those efforts were discontinued. The diffusion which occurred was among those hospitals which had adopted one lodge, and continued to develop new lodges without direct outside intervention. This finding is similar to a major finding of the previous implementation study (Fairweather, Sanders and Tornatzky, 1972) and indicates another variation of the "foot in the door" technique. It is the hospitals that are actively involved in aprogramy and which have ppeviously committed themselves to it, which continue to diffuse the innovation. Even in the YES/NO hospitals it was the more actively approached hospitals which diffused more. Just as with adoption, diffusion is more likely if the adopters are actively involved in the implementation process.- Diffusion Hypotheses In Table 9 we presented the cluster loadings which directly test each of the eight hypotheses for- warded earlier. Those loadings above .40 have been marked with an asterisk, and these indicate the vari- ables which are significantly related to diffusion.- We will now look at each hypothesis, and discuss any confirmation which appears in either the NO/NO, YES/NO or volunteer condition. 69 Hypothesis 1. All three conditions reveal variables indicating the importance of specialized roles for information- seeking. It is not mere appearance of such a group (as hypothesized) which is important, but certain character- istics of it. In the YES/NO and volunteer hospitals, diffusion is related to a higher status individual as head of the group, while in the NO/NO hospitals diffu- sion is related to a long; status person in that posi- tion. In addition, in the volunteer hospitals, the in- formal contact is of low status in the most diffusing hospitals. This latter finding may indicate that the formal group must be legitimized by a high status leader, though the "work" may be done by a more informal, low status contact. Hypothesis 2. Once again, all three conditions reveal signifi- cant loadings in this hypothesis. And once again, the NO/NO hospitals reverse the findings of the YES/NO and vOlunteer hospitals. The latter two groups associate nogg_money for staff rewards with diffusion while the former indicates igss_money for innovative idea rewards. This distinction between the-three conditions will be discussed further in the next section of this paper. Apparently, the amount of diffusion is not the only difference among them. 70 Hypothesis 3. There were no significant loadings indicating any relationship between diffusion and change in leader- ship. Hypothesis 4. Only two variables revealed a significant re- lationship between diffusion and crisis in the hospital. In the NO/NO hospitals the lower increase in budget (often seen as a crisis by our respondents) was related to diffusion. In the volunteer condition, the fact that there was less change due to a crisis is associated with diffusion, though mere occurrence of crises is not (as hypothesized). This is not a particularly strong con- firmation of our hypothesis. Hypothesis 5. Only in the volunteer hospitals are there sig- nificant relationships between power structure variables and diffusion. But these relationships are perhaps the most consistent and most confirming of the study. Diffu- sion is associated with less importance of influence (mean amount of influence) and, in particular, with less power in the_gonerally most influential disciplines (su- perintendent and psychiatry). This seems to associate less traditional and hierarchical structures with diffu- sion. 71 Hypothesis 6. There were no significant loadings indicating any relationship between diffusion and "pride." Hypothesis 7. Once again, only in the volunteer hospitals is i there a significant relationship between diffusion and a relevant variable. But in this case, the relation- ship does not support the hypothesis that diffusion is , positively related to a change-oriented hospital. Diffu- sion is significantly negatively related to the hospital wishing to start new programs. Apparently, the less a hospital is verbally committed to new programs, the more ‘E'diffusion there will be of any specific program. Perhaps acceptance of any new program is associated with those hospitals which are indiscriminate in their choice of new programs and not those who would carefully weigh a con- cept such as the lodge. Hypothesis 8. Only in the NO/NO hospitals is there anyindica- tion of the significance of systemic perspective. But this finding, in fact, is the opposite of the hypothesis. A concern with community needs is negatively related to diffusion. In the volunteer hospitals all of the systemic 72 variables are also negatively related and three approach significance. Such verbal systemic concerns are not re- lated to diffusion as hypothesized. Relationship Between Diffusion Cluster and Other Clusters In Tables 4, 6, and 8 we have looked at the re- lationship between the cluster domains found in each condition. Except for the NO/NO hospitals, diffusion is represented by an independent cluster, unrelated to other domains. Only in the NO/NO hospitals, which re- fused all attempts at intervention, is diffusion re- lated to other clusters: adoptiveness-innovativeness, expenses and superintendent influence. This, together with the finding that diffusion was related to more ac- tive persuasion attempts in the YES/NO hospitals, re- veals a great deal about diffusion as a concept and its relationship to outside intervention. Completely unplanned diffusion is related to more general hospital characteristics; in this case less power in the superintendent, lower increase in ex- penses and greater general adoptiveness-innovativeness. But once intervention is introduced into the system, that intervention is an important diffusion factor. In fact, diffusion is largely determined by those interven- tion attempts. Diffusion may be related to general Elgrt Ii~ 73 hospital characteristics if left alone, but not once in- tervention activities begin. The intervention attempts appear to alter other relationships which may exist with- out it. Overview of Additional Results In addition to the results which relate directly to diffusion, the cluster analyses employed disclose other interesting findings. The Non-Volunteers The clusters reported in Table 3 and Table 5 re- veal little difference between the 102 NO/NO and the 117 YES/NO hospitals. It would appear that the clusters pre- sented in Chapter III are consistent for hospitals which did not volunteer to implement the lodge. In Table 10 we present a comparison of these two populations. We have included the clusters of the implementation volun- teers for completeness. Eight of the nine YES/NO clus- ters also appeared in the NO/NO hospitals. The two con- ditions are virtually indistinguishable in terms of their cluster domains. Due to this similarity, much of the subsequent discussion will refer to these hospitals as implementation non-volunteers because the distinction between NO/NO and YES/NO hospitals is less important. In addition, this focuses further attention on the uniqueness of the volunteer hospitals. 74 :H dadom HoumsHo m op umHHEHm HH:MH£ no on HmuHucopH mH HoumsHo mHgH .muopmsHo HmpHmmon oz\oz ecu o .HoumsHo pmnu :H wovsHonH moHanum> Ho pop -85: ogu mucomoumon mama noumzHo some uoumm momonpcouwm :H Henson anew Hmv mswumonm so: How oouuHEEou .o HOV omsmcou HmuHmmom .m HOV ooaosHmeH Ho pedoe< .m HNO OmomHHo .O HNO momHHO .O panamaHmeH mo oesoe< .N HMO OOHNOOHEONOO .N HOV opcovaodeHoQSm .o HNV momcomxo HmpHmmom .o . Hmv mmoco>Hum>ocoH HNV nmomdomxo HmpHmmom m -mmo:o>Hum0d< .m nmv ammodo>Hpm>oddH - mmoco>Humod< .O HOV uaovcoudHquSm .O HNHV mmono>Huw>occH - mmoco>Humoo< .m HOV oNHm HmuHmmom .m Hmv momcoo HmuHmmoz .m HNNO OOHHHOOOO HNHO - mcHuocumm o.um:m 30Hoo modHHmHu Amy mucosHde Hmuou OOHOOENOHOH .N -mHO-OuaonHHOH HONOH .N oaHHOHOOHO OHOOHz: .N HOV OOHOOHHHO .H HOO HOOHOOHHHO .H HOV OOHOOHHHQ .H HoumsHu nonmaHu enoumsHu mm»\oz OOO mm»\mm» oz\mm» oz\oz .mHauHmmom uoouddHo> wee uooudsHo>-doz ooHududosonsH can :H mouamdoz wouopmsHu .OH oHan 75 For the non—volunteer hospitals, measures of the adoption of innovative programming formed reliable, rational clusters. The data indicates that the total number of new programs is related to a higher community locus score and/or a higher patient-autonomy score. This would indicate that if a hospital begins several programs, it is more likely that a community-orientation and patient autonomy will appear. This does not mean that all programs are innovative, only that there will be innovative elements appearing when many programs are on-going. Other than two committee variables in the YES/ No hospitals, the only variables included in the adop- tiveness-innovativeness clusters are those which define innovative programs. The hospital, demographic, atti- tude and power variables are not_inc1uded. In addition, this cluster is unrelated to other clusters which in- clude more general hospital variables. Such a finding forces us to consider the independence of adoptiveness- innovativeness as a concept. Up to this point we have been discussing this group of variables under the title adoptiveness-innova- tiveness. It is important to consider the necessity for such.afititle. From the perspective of many hospitals, the adoption ofgény program is an innovative act. How- ever, from the perspective of many theoreticians and 76 researchers, this is not the case. Rather, to them it is not innovative to adopt a program which does not in fact change the normal functioning of that institution in terms of role changes, status changes or actual or- ganizational changes. The concept we have empirically found crosses into both of those domains. It includes variables describing both quantitative and qualitative programming changes. Therefore the necessity of employ- ing the title, adoptiveness-innovativeness. In either case, it would appear that these con- cepts are associated with variables not tapped in this study. This suggests that for many institutions, the spreading of an innovation must be designed with each institutions' characteristics well in mind. It is not the general hospital variables that are important, but variables specific to each hospital. Approaches to each hospital must be designed with both the innovation and the institution in mind. The findings of the hos- pital implementation study reveal that for the iodggj certain approaches were appropriate, and seem most ef- fective in spreading it as an innovation. This finding is not contradicted here, though the generalizability of previous results are limited by these findings. The independence of our measures of general adoptiveness-innovativeness of programs is an important finding. We see no support for the concept of the 77 generally adopting and innovative hospital. For these 219 hospitals, and to a limited extend for the 25 vol- unteers, adoption of innovative programs could not be predicted from the general hospital variables measured. Adoption and innovation appear to be independent and unique functions. They are most likely associated with characteristics unique to each institution, each inno- vation and with general aspects of the innovation pro- cess itself which were not tapped in this study. With regard to power structure, we find essen- tially two levels of influence for new programs, super- intendent and "others." Psychiatrist influence seems to cross both lines while not directly aligning with either. The most striking result is the flatness of power of the middle disciplines: psychology, social work, nursing and vocational rehabilitation. The even- ness may be due in part to the unitization movement in many hospitals. Many respondents indicated that divid- ing the hospital into units, levels out the influence of the hospital structure. However, such flatness is on the unit level among middle disciplines. The super- intendent is still the most powerful, and the psychia- trist still exhibits his influence. Table 11 gives the mean values of amount of influence and supports the above discussion. 78 Table 11. Mean Values of Amount of Influence - Implementation Non—Volunteers.a NO/NO Hospitals YES/NO Hospitals Superintendent 4.53 Superintendent 4.39 Psychiatry 3.78 Psychiatry 3.80 Nursing 3.47 Psychology 3.47 Psychology 3.35 Social Work 3.34 Social Work 3.35 Nursing 3.27 aRanked such that ”l" is no in- fluence - "5" is greatest influence. Many assumptions regarding innovation and diffu— sion revolve around the importance of budget, staff and census considerations. We find no basis for this con- cern in our sample. Hospital expenses and census data were related only to themselves, not to innovativeness or any other cluster.. A similar finding with regard to lodge adoption was found in the previous implementa- tion study. The Implementation Volunteer Hospitals For numerous reasons, the-25 implementation vol- unteers were examined separately in the present study. They have each been subjected to active implementation 79 attempts, they have diffused more information about the lodge and their cluster analysis revealed different do- mains from those found in the non-volunteers. We must also reemphasize that the reduced N size has serious implications for interpretability of our analysis. As noted above, adoptiveness-innovativeness is a separate and distinct cluster from diffusion, unlike thefinding in the non-volunteer clusters. For these hospitals, it is significantly related to other types of variables. The superintendent's being innovative and from inside the hospital are related to innovative- ness. Carlson (1965) had indicated that change in leadership, less tenure of leadership and superinten- dent's frompoutside the school system were related to innovativeness. "These results stand in opposition to his findings. Also, variables related to the search for new programs are related to adoptiveness-innovativeness in a curious way. We find here variables which seem to indicate a fluid, less structured new program committee. It meets rarely, does little with the information gathered, and the source of information is seen as an informal, low-status individual. In addition, there is little money allocated for a library fund. Thus adoptiveness-innovativeness is associated with innovative 80 superintendents who come from within the hospital system, and with new program sources which are informal, low sta- tus and generally not rigidly structured. The positive information gathering variables which were hypothesized to be associated with diffusion, are in fact associated with stability. Positive informa- tion gathering variables such as more funds for workshops and travel, and a larger new program committee, are asso- ciated with variables which describe a stable, established hospital; few crises, satisfaction with programs, little turnover of staff and generally a great amount of influence for the middle disciplines. The new program subunit (_y- pothesis 1) is a sign not of innovativenessJ but stability. It seems that once such a committee is established and part of the hospital's status quo, it reinforces the stability which that hospital may have gained. New pro- grams may be looked at by such committees, but new pro- grams do not appear in_that hospital. The committee may relieve the hospital of the responsibility for actually implementing new, innovative programs. Innovative pro- gramming occurs in those hospitals which place new pro- gram responsibility in less well-established hands. A stable hospital is not threatened by a committee looking for new programs, but it will also not change and will not be associated with general innovativeness. In our 25 volunteers, stability is not associated with 81 diffusion of a specific innovation nor general innova- tiveness. 0n the other hand, diffusion is not directly associated with a crisis-orientation as proposed by Schon (1967). Interpretations and Conclusions Limitations As is typically the case in non-laboratory re- search, the most serious deficiencies of the present study involve variables uncontrollable by the research team. The most serious limitation in terms of hypothesis- testing lies in the lack of diffusion of our innovation. Although this is in itself an important finding, it cre- ates difficulty for any hypothesis-testing as mentioned earlier. It is difficult to test hypotheses about an event which only rarely occurs. To useRogers' (1971) terms, our innovation lacks relatiVe advantage, is too complex, incompatible, unobservable or untriable to be adopted by a majority of our population. Another test of diffusion which involved a more easily spread innovation might have given us further insight into the diffusion process. Our findings can only apply to Complex programs like the lodge. Other limitations involve measurement and research design. Ideally we would have employed face-to-face 82 interviews, site visits, direct measurement of innova- tive programs, and the like. But time and resource limitations precluded the ideal. In light of the number of personnel changes which we found, and the fluid nature of many of the institutions interviewed, tracing down our preferred respondents was also difficult. Our hierarchy of pre- ferred respondents may not be ideal, and arguments could be made for other choices. After extensive dis- cussion with the entire research team, the present hierarchy was chosen. The fact that neither status of respondent nor past involvement of respondent appeared as important variables, lends some confirmation to the validity of our decision. A somewhat more serious limitation, which has been discussed previously, involves the role of our interviewers. One interviewer did evoke significantly different responses on some questions than the other two. It is unlikely that further training would have eliminated this bias. Rather, it appears that such differences were due to the personality of that one interviewer. The fact that his influence was essen- tially inconsequential to the cluster analysis, allows more confidence in our findings. We suggest, however, that such a variable be carefully controlled in-future diffusion studies. brie empi I21 But the si: pi‘ nu te ni Hc If“. 83 Finally, the question of sample size deserves brief discussion. Considering the fact that we have employed virtually the entire pOpulation of State and V.A. mental hospitals, this should not be a concern. But, the most significant associative findings are in the implementation volunteer hospitals, where the N size is only 25. Our results for the other 219 hos- ! 1 I ..m _ ...‘6. pitals are not limited by this concern, but N size must be considered in any interpretation of the volun- teer hospitals. The limitations discussed above must be recog- nized in any interpretation of the results of the study. However, the results obtained cannot be discounted or ignored because of these limitations. Planned and Unplanned Diffusion Much of the theoretical background for the pres- ent study is based upon the work of Rogers (1962, 1971) and Havelock (1971). Rogers' conceptualization of the social interaction model is founded in large part on examples of unplanned diffusion. We must recognize, however, that there are important differences between adoption of agricultural implements, prescription drugs, or contraceptives and the adoption of a complex, social system change, such as the lodge. For one, lodge adop- tion involves role changes inconsistent with past 84 behavior. Adapters are not merely accepting a better way to grow corn, but rather a new life style for them- selves and their clients. In addition, an important element of Rogers' model is the interaction which exists between social system elements. Such interaction will occur in a small Iowa community; but we have studied a population without such effective communication channels. It is a large system which encourages only infrequent commun- ication between elements. As we will show in the next section, the communication which does occur even in the best circumstances is haphazard and leads to little more than discussion of the concept. Therefore, the lack of unplanned diffusion of the lodge concept must be examined in the light of such limitations. The fact that some diffusion did occur in those hospitals where active implementation attempts gong made is further evidence of the importance of plan- ning any innovation attempt. Future research should in- vestigate the best methods for planning such diffusion, just as Fairweather, Sanders and Tornatzky, (1972) have done in the hospital implementation study. An encouraging aspect of the present study was the unplanned dissemination of lodge information which emanated from at least one of the volunteer hospitals. As a direct consequence of implementing the lodge, 85 several members of this hospital's planning group con- tinued to spread the lodge concept throughout their home state.- They presented the information to state psychological, social work, and vocational rehabilita- tion meetings. A psychologist visited other state hospitals and even brought staff members from those hospitals to the lodge. The mean diffusion score for this state is 4.3, far above the national average. This hospital suggests the value of a "diffu- sion center." Information about that center's innova- tion can easily be spread. The innovators can capital- ize on the ease with which such local centers can spread information. But, as we can see, although information is spread, actual adoption is not necessarily forth- coming. Such diffusion centers would have to add an ac- tive, implementatiOn attempt to get outstatehospitals to go beyond mere discussion. Future Research Several possibilities for future research have been mentioned. The present study raises some inter- esting methodological questions which could be answered by future research. Basically, these revolve around the question of finding the most effective means of gathering diffusion data from a national sample. One could compare phone vs. written techniques, superintendent h h 14.4 86 vs. "other" respondents and perhaps attempt to gather more behavioral measures of diffusion. It would also seem beneficial to build diffusion information directly into the implementation part of experimental social in- novative research. More accurate information would be available if it were gathered during the actual diffu- sion period, rather than depending upon recall. Questions surrounding the type of innovation investigated might also be forwarded. As in Carlson's (1965) study of educational change, categorizing inno- vations by Rogers' (1971) criteria of relative advan- tage, complexity, trialability, observability and com- patibility would be helpful. In this way one could compare important variables in both complex and simple changes. The importance of the active change agent has been further highlighted in this study. The diffusion which occurred, occurred in those hospitals which had become active in the change process; and diffusion did not occur in those hospitals untouched by planned change attempts. Further research must be done to investigate the role of the change agent, and in particular, the possibility of an adopting institution becoming a diffu- sion center. A diffusion study could be developed to evaluate the effectiveness and feasibility of creating such.centers. A major research effort should be designed 9-7 to CI Ties spon adop 87 to compare different methods for utilizing such centers. These could be statewide or regional. They could be re- sponsible for spreading an innovation which they have adopted. BIBLIOGRAPHY 1'- BIBLIOGRAPHY Allen, T. Managing the Flow of Scientific and Techno- logical Information, Ph.D. Thesis, Cambridge, Mass., Massachusetts Institute of Technology, September, 1966, cited in Havelock, 1971. Beal, G. M., Rogers, E., Bohlen, J. M. "Validity of the Concept of Stages in the Ad0ption Process," ‘ Rural Sociology, June, 1957, Vol. 22, Pp. 166-168. Becker, A., Murphy, N. M., Greenblatt, M. "Recent Ad- vances in Community Psychiatry," in Bindman, et gi., Perspectives in Community Mental Health, CEicago: Aldine Publishing, 1969. Belknap, 1. Human Problems of a State Mental Hospital, New Y6ik: McGraw-Hill, 1956. Bindman, A., and Spiegel, A. Perspectives on Community Mental Health, Chicago: Aldine Publi§hing, 1969. Brickell, H. M. "State Organization for Educational Change: A Case Study and Proposal," in Miles, M. F. (Ed.), Innovation in Education, New York: Bureau of Publications, TeachErs College, Columbia University, 1964. Carlson, R. Adoption of Educational Innovations, Eugene, Oregon: University of Oregon, I965. Carlson, R. Executive Succession and Organizationgi Change, Chicago: University ofChicago, 1962. Coleman, J., Katz, E., and Menzel, H. Medical Innovation: A Diffusion Study, New York: B6bbs-Merrill, 1966. Fairweather, G. W. Social Change: The Challenge to Sur- vival, New Jersey: IGEneral Learning Press, 1972. Fairweather, G. W. Transcending Revolution, Unpublished manuscript, 1972. Fairweather, G. W. (Ed.), Social Psychology in Treating Mental Illness: aniExperimental Approach, New York: John Wiley and Sons, Inc., 1964. 88 89 Fairweather, G. W., Sanders, D. H., and Tornatzky, L. G. Changing Mental Hospitals, (1972, in press). Fairweather, G. W., Sanders, D. H., Cressler, D., and Maynard, H. Community Life for the Mentally Ill, Chicago: Aldine PuhliShing Co., 1969. Fairweather, G. W., Simon, R., Gebhard, M., Weingarten, E., Holland, J., Sanders, R., Stone, G., and Reahl, J. "Relative Effectiveness of Psychotherapeutic-Pro- grams: A Multicriteria Comparison of Four Programs for Three Different Patient Groups," Psychological Monographs, Vol. 74, No. 5, 1960. Glaser, E. Utilization of Applicable Research and Demon- stration Results, Dept. of H.E.W., Project RD-1263 G,lMarEh, I967. Goffman, E. Asylums, New York: Anchor Books, 1962. Griffiths, D. "Administrative Theory and Change in Organ- izations," in Miles, M., (ed.) Innovation in Educa- tion, N.Y., Bureau of Publications, Teachers CETIege, Columbia University, 1964. Guba, E. "The Change Continuum and Its Relation to the Illinois Plan for Progravaevelopment for Gifted Children," Paper delivered to a Conference on Educational Change, Urbana, Illinois, March, 1966, cited in.Havelock, 1971. Havelock, R. Plannin for Innovation through Dissemina- tion and Utilization of Knowled‘e, nn Afbor, Institute for Social Research, 71. Havelock, R., and Benne, K. "An Exploratory Study of Knowledge Utilization,” in Watson, G. (ed.), Concepts for Social Chan e, Baltimore, Md., Moran Printing Service, u lished by N.T.L., NEA for COPED, March, 1967. Hobbs, N. "Mental Health's Third Revolution," in Bindman and Spiegel, Perspectives on Community Mental Health, Chicago: Aldine Publishing, 1969} ”Hospitals" Journal of the American Hospital Association, August,'1970. "Hospitals" Journal of the American Hospital Association, August, 1971. 90 Katz, D., gt ni. Studies of Innovation and of Communica- tion to the Public, Stanford University, Schhol of Communication Research, 1962. Katz, D., and Kahn, R. The Social Psycholo y of Organ- izations, New York: John Wiley an Sons, Inc., Kesey, K. One Flew Over the Cuckoo's Nest, N.Y., Viking Press, 1962. Knoerr, A. "The Role of Literature in the Diffusion of Technological Change," Special Libraries, May- June, 1963, Vol. 54, pp. 271-275. Lewin, K. Field Theory in Social Science, N.Y., Harper and Row, 1951} Lewin, K. "Group Decision and Social Change," in Swanson, G. E., et gi., Readin s in Social Ps - chology, Henry Hart and Co., 1852, pp. 459-473. Lawrence, P. R., and Lorsch, J. W. giganization and 1 Environment, Homewood, 111., chard D. Irwin, Ihc., 1969. Lippitt, R., Watson, J., and Westley, B. The Dynamics of Planned Change, N.Y.: Harcourt, Brace and Co., Inc., 1958. Mann, F. C. and Williams, L. K. "Observations on the Dynamics of a Change to Electronic Data Process- ing Equipment," Administrative Science Quarterly, Sept., 1960, Vol. 5, pp. 217-256. Marrow, A., Bowers, D., and Seashore, S. Management by Participation: Creatin a Climate for Personal and Organizational’Deve opment, N.Y.: Harper and Row, 1967. Michael, D. Factors Inhibiting and Facilitatin the Acceptance of Educational’lnnovations, ashington, D7C., Institute for Social Policy Studies, 1965, (mimeo, cited in Havelock, 1971). Miles, M., and Lake, D. "Self-Renewal in School Systems: A Strategy for Planned Change," in Watson, G. (ed.), Concepts for Social Chan e, Baltimore, Md., Moran Printing Service, PubliE ed for N.T.L., NEA for COPED, March, 1967, pp. 81-88. 91 Pasamanick, B., Scarpitti, F., and Dinitz, S. Schizo- phrenics in the Community, N.Y.: Meredith P 5— lishing C6.,‘1967. President's Conference on Technical-Distribution Research f6r the Benefit of Small Businesses, Wishington, ' D.C., Office of‘Technical Services, U.S. Depart- ment of Commerce, Sept., 23-25, 1957, cited in Havelock, 1971. Rogers, E. Communication of Innovations, N.Y., The Free Press of Glencoe, Inc., 1971. Rogers, E. Diffusion of Innovations, N.Y., The Free Press oflGlencoe, Inc., 1962. Ryan, B., and Gross, N. "The Diffusion of Hybrid Seed Corn in Two Iowa Communities," Rural Sociology, March, 1943, Vol. 8, pp. 15-24. Schein and Bennis. Personal and Organizational Chan e Throu h GroupiMethods, N.Y.: John Wiley an Sons, nc., 1965. Schon, D. Technology and Change, N.Y., Delacorte Press, 1967. Siegel, S. Non-Parametric Statistics for the Behavioral Sciences, N.Y.: McGraw-Hill Boole5., I956. Swanson, G. E., et El: Readin s in Social Ps cholo , Henry HOT? and Co., I952, pp. 459-473. Thelen, H. A. "Concepts for Collaborative Action In- quiry," in Watson, G. (ed.), Concepts for Social Chan e, Baltimore, Md., Moran Printing Service, PuEIished by N.T.L., NEA for COPED, March, 1967. Tryon, R. and Bailey, D. Cluster Analysis, N.Y.: McGraw- Hill Book Co., 19703 Veyette, J. H., Jr. ”Planning for Data Retrieval," Automation, Jan., 1962, Vol. 9, pp. 129-132. Watson, G. (ed.) Concepts for Social Chan e, Baltimore, Md., Moran Printing Service, Publis ed by N.T.L, National Educational Association, for COPED, March, 1967. 92 watson, G. Social Psychology; Issues and‘Insights, Philadelphia, Pa., 1ppincott,1966. Wilkening, E. "The Communication of Ideas on Innova- tion in Agriculture," in Katz, et al., Studies of Innovation and of CommunicatiEh—fo the PuBIic, SfanfordlU., School for Communicatihn Résearch, 1962, pp. 39-60. APPENDICES 93 APPENDIX A Phone Questionnaire Introduction "Hello. I'm Dr. and I am a colleague of Dr. George W. Fairweather at Michigan State University. Dr. Fairweather and (previous research contact) have been T“‘ involved for a number of years in research supported by + NIMH research grants. Our present research is concerned . . with information which you might have about the mental health programs in your hospital. We are hoping to gain 53; some insight into programs which hospitals find work for them. I would like to ask you some questions about your programs and also about your hospital in general. All information will be kept confidential and we will not use your hospital's name in any way. It will only take 15-20 minutes, and will help us a great deal. Could you help me?" Yes No . IF "NO" ASK: "Could you suggest someone who might be able to help us?" (Ask lst contact to switch you to operator and ask for 2nd person on list. Continue down list until a "Yes." If the last attempt also says "No," try person suggested above. IF "YES," continue questionnaire. 94 APPENDIX B Questionnaire Items Used to Measure Degree of Adoption of the Community Lodge (Diffu) 1. Knowledge a. Have you heard of the Fairweather Community Lodge Program for schizophrenics? IF "YES," read the following statement and go to question b. IF "NO," read following statement and ask if the respondent now remembers having heard of the lodge. If a second "NO," go to hypothesis-testing questions. In (month, year) our research team contacted you about such a program. Subsequently, your hospital was contacted by (implementation consultant). We later inquired whether or not you would like to set up the Community Lodge Program for (name of hospital). That was in (month, year) and your hos- pital did (not) want to set up such a program. b. Please describe the lodge program as you remember it? 2. Persuasion We are now interested in your hospital's subsequent reaction to the community lodge program information which you received. The following questions refer 9S specifically to the time after we last contacted you for a decision, that is, after (month, year). a. Did anyone in the hospital attempt to get more information about the lodge? b. Are you aware of eny discussion about the lodge program which has occurred since (implementa- tion decision date)? IF "YES," TO b.l_GO TO c. c. How many such discussions were there? d. Who was discussing it? Number of people? Which disciplines? e. In what context were they discussing it? 3. Decision a. Was a formal decision made about adOpting the lodge program since the (implementation decision date)? b. What was that decision? c. Who had the greatest amount of influence in that decision? d. What action was taken subsequent to that decision? 4. Confirmation (Action-adoption) a. Did the lodge program become a part of your hos- pital's activities? b. Please describe the operation of the lodge fully. c. When did it begin operating? 96 Is it still in operation? If, "NO," When did it discontinue? Why was it discontinued? Have more than one lodge begun? How many? Was information about the lodge requested by others outside of your hospital? Have you or any of the staff sent information to others about the lodge program? Please des- cribe. 97 APPENDIX C Questionnaire Items Used to Measure Hypotheses l. Specialized subunit for seeking_newprogram information. a. Is there a group or individual whose function is to look for mental health treatment programs new to your hospital? IF INDIVIDUAL b. C. What is his status? What percent of his time is spent in this func- tion? d. What does he do with the information he gets? IF GROUP e. What is the status of the head of that group? R. Is What percent of their time do most members of the committee spend in this function? How many people are there on that committee? Which disciplines are represented? How often does the committee meet? . What do they do with the information they get? thereanyone who is an informal contact for new programs? 1. IF "YES," what is his status? 2. Goals of the hospital emphasize seeking_new_program infhrmation. a. Is reporting and receiving information about new programs a goal of your hospital? b. Do staff members of your hospital present their new programs to others in your hospital, as at workshops, in-hospital training, etc.? 98' c. Are hospital funds available for any of the fol— lowing activities? (1) a library fund? (2) travel funds to attend workshops and con- ventions? (3) rewards to personnel for new program ideas? (4) in-hospital workshops for visiting profes- sionals? d. What is given weight in evaluating you by your superiors? This was measured by the following five—point scale items. How important is (l) (8) administrative ability ability to do therapy research ability supervision of subordinates attending meetings innovative ideas of treatment ability to get patients out of the hospital ability to get along with co-workers Item 6. was the.key item. In order to place its importance in perspective, this variable was scored by giving one point for each item that number 6. was more important than and one-half point for ties. The higher the resulting score, the more important innovative ideas are in the respondent's evaluation as compared to other activities. 3. Change in leadership - a. Have any new administrators joined your hospital since our last contact? b. How many? What positions? What date? c. When did the present superintendent of your hospital become superintendent? c. Did the present superintendent come from within the hospital's personnel or from outside the hospital? 4. Crisis situation a. Has any kind of major crisis occurred in your hos- pital since we last contacted you? 99 IF YES b. Please describe. How many such crises? c. When did that occur? d. How well was that crisis resolved? (5 point scale) e. Was there a great change in the hospital's functioning due to the crisis? Fflg (5 point scale) 5. Pattern of influence in decision-making about new pro- grams . a. In general, how much say or influence do you feel each of the following units has on development of new treatment programs? Please rate from one to five with one being "no influence," and five being .Hr "a very great deal of influence." Amount of influence Unit Initiate Approve Implement Resist a. Supt. b. Psychology c. Nursing d. Voc. Rehab. e. Social Work f. Psychiatry For each of the above groups, please give me the fol- lowing information? Is the influence of the (read unit names, changing order each time) in terms of which of the following? Do these units initiate new programs, approve new programs, implement new programs or generally resist new programs? (Check___ above) In order to measure "breadth of influence," a score was developed which simply counted the number of areas of influence checked by the respondent. "Total" influence was calculated by multiplying amount of influence and number of areas of influence together for each discipline. b. 100 Please describe the superintendent's role in your hospital. 6. Pride a . We need to obtain a subjective assessment of your treatment facilities as they relate to other mental hospitals in the country. Equat- ing 100% to the best facility we would like you to indicate what percent of this optimal per- formance you personally feel your hospital is achieving? 0-208 1 21-40% 2 41-60% 3 61-80% 4 81-1008 5 b. How much better do you feel your programs should be? (5 point scale) 7. Change orientation a. How does your hospital generally feel about start- ing new programs? (5 point scale) . Do you know of any programs which have been approved by the hospital, but not implemented in the last 5 years? How many? . Do you know of any programs which have been discon- tinued in the last five years? How many? 8. Systemic perspective This variable was measured by the responses to the fol- lowing S-point scale items. To what extent are the following concerns expressed by by leadership of your hospital? (1) relationship of hospital to local community (2) search for the optimal answer to mental health care (3) evaluation of present services (4) responding to community needs (5) active search for new programs Knowledge Persuasion Decision Confirmation 101 APPENDIX D Diffusion-Adoption Stages (Diffu) l...Never heard of the lodge, even though they may have heard of Fairweather's name. Can describe nothing about the lodge. 2...Heard of the lodge and can describe it, even superficially; no subsequent dis- cussion in or out of the hospital.‘ 3...Knows of minimal discussion about the lodge since our last contact. Evidence that less than five persons have dis- cussed the lodge. 4...Great deal of discussion in the hospital since our last contact, either informal or formal. 5 or more persons involved in discussions, greater than 10 discus- sions about it. 5...Decision since our visit or last phone contact. Either lodge turned down or no decision arrived at, though brought before appropriate persons. Decision on ward program only if it is specifi- cally intended to be pre-lodge. 6...Lodge (or pre-lodge ward) decision was yes, regardless of subsequent action. Little or no action towards lodge set- up, though there may have been some further discussion. 7...Pre-1odge ward or "semi-lodge" set up with intention of moving to full lodge. If ward set up, but using for purposes other than to feed lodge, score a 6. "Semi-lodge" must be based on the lodge sources, and may include 1/4 way house if it is basically a lodge on the grounds, or a lodge off the grounds with staff con- tact and direction greater than in lodge piototype. Action is based on hospital's owledge of the Fairweather lodge. 102 8...Lodge set up on the lines of the Palo Alto model, particularly regarding no live-in staff, and some degree of work in the lodge. Must have functioned for 6 mos. or still be in operation. If discontinued after less than 6 months, give it a 7. If no work, give it a 7. note: A hospital is to be scored on the highest level reached during the time period from implementation decision date to the present contact. For example, if the respondent can describe the lodge at the present time, and no other activity about the lodge occurred during the time period, then the hospital would receive a 2. If the contact claims he once knew what the lodge was, but cannot now describe it, the hospital would receive a l. Persuasion, decision and confirmation activity described must have occurred during the proper time period. Diffusion Score for 25 Volunteer Hospitals (Diffu 2) l...No additional discussion about movement towards the lodge. 2...Some discussion of lodge development, continuation of ac- tion started with researcher's help, but no new action. 3...Actua1 new action towards development of the lodge. 103 For those hospitals in the middle of adopting a first lodge at the last contact this is a measure of further progress. For those hospitals which had adopted a lodge by the last contact this is a measure of progress towards adoption of a second lodge. In both cases, it is again, movement towards adoption of a lodge since our last contact. In addition, Diffu is recorded for the volunteer hospitals. As a measure of progress towards adoption it H. I ' ‘ ‘1 is confounded with implementation attempts and is essen- tially a change score, but not a diffusion score. H .. 's' N ' 1 when“ “1 104 APPENDIX E Community-locus Score definition: Community involvement of the hospital in the treatment of its patients. How much is the hospital willing to involve itself in commun- ity experience of patients or ex-patients. l...Totally in-hospital orientation with no concern for patients community experience. 2...Simulation of community experience in the hospital. Training for community living without entering the community in a living or work situation, and without formal arrangement with community employer. In- hospital workshop included here, for in-patients. Include family therapy in hospital. 3...Helps patient in his community experience without structuring that experience. Includes follow-up, after care in which staff is available for follow- up, but concerned little with the structure of the patient's total community experience. Includes day hospital, since assumption is that the hospital ex- perience is necessary for community adjustment, though the hospital does not attempt to help with actual life of the patient in the community. Formal relationship for feeding patients to community services. 4...Structures some of patient's time in the community, trains community living in the community, patient‘ spends most of his waking time in the community. Sheltered workshop in the community or 1/4 way house with work in the community. 5...Hospita1 takes responsibility for patient's time in the community and.structures that time in the community. All of the patient's time in the community situation which is run by the hospital, or that the hospital takes responsibility for (1/2 way house, boarding house, foste§ care, if directly the responsibility of the hos- pital.. 105 APPENDIX F Autonomy Score 1...Staff living with, or attendant to patients at all times. Patient's "home" supervised by staff or pseudo-staff (foster care parents, 1/2 way house mothers). Includes traditional wards, with full time staff. r11 2...Staff available basically as consultants, but on a 1 daily basis. Staff does not supervise all activities of patients, but rather give patient some responsi- bility for making his own decisions as with patient self-government. Includes day care - daily contact without staff living with patient. Waking hours in- j L clude both supervision and experience in peer run situation. Day hours may include supervision while night is free of staff intervention, or vice versa. w 3...Staff available only on an infrequent and probably irregular basis. No staff actually living with patient, peer-run situation; living at home or alone with follow-up program. Living situation does not include staff members being present at all times; 1/4 or 1/2 way house without live-in staff, no staff attendant during day or night hours. HICHIGAN STATE UNIV. LIaRaRIEs IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 31293100158421