AM EXPLORATQRY STUDY COMPARING CHARACTERISTICS. OF INNOVATORS AND NONrINNOI/ATORS I AT A LARGE UNIVERSITY ‘ Dissertation for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY STEVEN GREGORY SACHS 1976 IIIIIIIIIIIIII COMPARING CHARACTERISTICS OF 5 INNOVATORS AND NON-INNOVATORS AT A LARGE UNIVERSITY .. presented by Steven Gregory Sachs has been accepted towards fulfillment of the requirements for Secondary Education Ph' D‘ 9 degree in and Curriculum florid... Major professor Date « June 30, 1976 0-7 639 a- sarcoma Iv - , I. Ill!“ & SINS' ‘ II am amnm Inc. 3 _ may 3"!ng MSU LIBRARIES RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. 89 “'3‘38 mm M I” . m . W fire/772,4- J/x/t.) ABSTRACT AN EXPLORATORY STUDY COMPARING CHARACTERISTICS OF INNOVATORS AND NON-INNOVATORS AT A LARGE UNIVERSITY By Steven Gregory Sachs This study had two purposes. The first purpose was to identify in a higher education setting, specific characteristics of innovators and their departments which differed from non-innovators and their departments. The second purpose was to validate part of Rogers' and Shoemaker's model which deals with the characteristics of adopters of innovations. Only 18 of the 45 variables from this model were selected for this study. The variables used in this study were age, status, size of teaching load, specialization of teaching responsibility, fatalism, innovativeness, social participation with departmental colleagues, integration with their social system, cosmopoliteness, information seeking about instruction, knowledge about instructional innovations, opinion leadership, membership in modern and/or integrated system, norms on the importance of teaching, norms on innovativeness, norms an instructional strategies, resources for instructional improvement, and stability of instructional assignments. These 18 variables were represented by 27 different measures. To identify the differences between innovators and non- innovators, this study used an ex post facto multivariate comparison design. The data were collected through questionnaires distributed Steven Gregory Sachs to 345 faculty at Michigan State University. These faculty were divided into three groups for the purposes of analysis. The first group was composed of faculty who had received grants from the MSU Educational Deve10pment Program (EDP) for an instructional development project for improvement to an undergraduate course. These were the instructional innovators. The second group was composed of faculty who reported that they had not been involved in an instructional deve10pment project during the preceding five years. These were the non-innovators. The third group was composed of faculty who reported that they had been involved in an instructional development project during the past five years, but who had not received EDP support. These were the unsupported innovators. The data were analyzed by two multiple regression techniques: simple multiple regression and discriminant function analysis. This analysis occured in four parts. First, a regression equation to predict the number of projects an innovator had done was generated from data supplied by the EDP project directors. Second, the same data that were used to generate the first regression equation were used to generate a second regression equation which was to predict a combined measure of success for the innovator's first project. Both of these equations were then analyzed using discriminant function analysis to determine whether they could identify variables which would differentiate between innovators and non-innovators. The third part in the data analysis was the use of a separate discriminant function analysis to identify variables which differentiated between the known innovators (EDP project directors) and the known non- innovators. Finally, the data from the unsupported innovators were ‘g-C Steven Gregory Sachs compared with the data from the innovators and the non-innovators. This four-part analysis yielded information about which variables, acting individually or in combination, were significantly different between innovators and non-innovators, and also provided linear equations for use in future analysis of individual faculty members. 0f the 18 variables included in this study, only eight were found to be significantly different between innovators and non— innovators. These were: innovativeness, integration with the social system, cosmOpoliteness, information seeking about instruction, opinion leadership, norms on innovativeness, norms on instructional strategies, and stability of instructional assignments. Major findings. The data analysis provided the following major findings: 1. Innovators differed from non-innovators by showing: a. Greater innovativeness b. More integration with the social system c. Greater cosmopoliteness d. Greater information seeking about instruction e. Greater opinion leadership change 2. Innovators' perceptions of their departments differed from the perceptions of non-innovators by showing: 3. Less supportive norms on innovativeness b. Less supportive norms on instructional strategies c. More stability of instructional assignments An additional major finding was that there were three distinct groups of faculty: innovators, early adopters, and non-innovators. As a result of these findings, four conclusions were reached. Steven Gregory Sachs These conclusions were: 1. Eight differences can be identified between innovators and non-innovators in terms of personal and social system characteristics. Three groups of faculty can be identified: innovators, early adopters, and innovators. Innovators perceive that their departments do not provide sufficient financial and/or psychological support for instructional innovation. Only a portion of the individual and social system characteristics from the Rogers and Shoemaker model on the diffusion and adoption of innovations appears generalizable to higher education settings. AN EXPLORATORY STUDY COMPARING CHARACTERISTICS OF INNOVATORS AND NON-INNOVATORS AT A LARGE UNIVERSITY By Steven Gregory Sachs A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Secondary Education and Curriculum Instructional Development and Technology 1976 © Copyright by STEVEN GREGORY SACHS 1976 ACKNOWLEDGEMENTS I want to recognize a number of people for the parts they played in helping me complete this dissertation. Their efforts made it possible for me to tackle a project of this size and complexity. I also want to give credit to the Educational Deve10pment Program at Michigan State University for its financial support of my research. Special thanks go to the members of my doctoral committee. I owe a particular debt of gratitude to Dr. Allan J. Abedor, my dissertation director, for the countless hours he spent reviewing, guiding, and encouraging my dissertation. In addition I want to thank the other members of my committee--Dr. Bruce Miles, Dr. Erwin Bettinghaus, and Dr. Stephen L. Yelon--for giving me the freedom to undertake this project and for continually demonstrating their confidence in me. While not officially members of my committee, Dr. Robert Davis, Dr. Lawrence Alexander, and Dr. Paul W.F. Witt also contributed important suggestions and encouragement throughout my research. Special credit must also be given to Mary Shepard and Pat Sweet for handling so many of the clerical details and the voluminous typing associated with this project. Most important of all, however, I want to recognize the tremendous contribution made by my wife, Betsy. Without her love, encouragement, support, and sacrifice, this entire effort would never have gotten beyond the idea stage. ii TABLE OF CONTENTS LIST OF TABLES. . . . . . . . . . . . . . . . LIST OF APPENDICES. . . . . . . . . . . . . . Chapter I THE PROBLMO O O O O O O O O O O O O 0 Introduction. . . . . . . . . . . . Purpose . . . . . . . . . . . . . . Importance. . . . . . . . . . . . . Generalizability. . . . . . . . . . Underlying Theory . . . . . . . . . Variable Selection. . . . . . . . . The Research Questions. . . . . . . Innovators vs. Early Adapters. . . Definitions of Terms. . . . . . . . Overview of the Study . . . . . . . II REVIEW OF RELATED LITERATURE . . . . . Introduction. . . . . . . . . . . . Overview of the Literature. . . . . Organization of the Review. . Studies of Adapters and Nonradopters. . . Studies of Attitudes Toward Innovation. . Other Studies . . . . . . . . . . . Comparison with Findings from Other Socioeconomic Variables. . . . . Personality Variables. . . . . . Communication Behavior . . . . . Social System Variables. . . . . Summary . . . . . . . . . . . . Implications for this Study . . . . III DESIGN OF THE STUDY. . . . . . . . . . Introduction. . . . . . . . . . . . The Research Design . . . . . . . . Research Questions. . . . . . . . . . Operational Definitions . . . . . . Population and Sample . . . . . . . Instrumentation . . . . . . . . . . Questionnaire Development and Distribution . It” selection I C O O I O O O 0 Fields Reliability and Validity. . . . . . . . . . iii vi vii Page H 14 15 17 18 19 21 24 24 25 26 28 28 29 30 30 33 , 34/ 35 39 41 41 42 44 Chapter IV V Page Data Analysis . . . . . . . . . . . . . . . . . . . . . . 45 Overview . . . . . . . . . . . . . . . . . . . . . . . 45 Development of Multiple Regression Equations . . . . . 45 Direct Comparison Using Discriminant Function and Analysis . . . . . . . . . . . . . . . . . . . . . 47 Summary . . . . . . . . . . . . . . . . . . . . . . . . . 43 Multiple Regression Analysis. . . . . . . . . . . . . . . 49 Significance Tests . . . . . . . . . . . . . . . . . . 50 Inclusion of Variables . . . . . . . . . . . . . . . . 51 Replacement of Missing Data. . . . . . . . . . . . . . 51 Population Variance. . . . . . . . . . . . . . . . . . 52 Discriminant Function Analysis. . . . . . . . . . . . . . 52 Discriminant Functions and Coefficients. . . . . . . . 53 The Classification Process . . . . . . . . . . . . . . 53 Significance Tests . . . . . . . . . . . . . . . . . . 53 Inclusion of Variables . . . . . . . . . . . . . . . . 54 Replacement of Missing Data . . . . . . . . . . . . . 55 Population Variance. . . . . . . . . . . . . . . . . . 55 RESULTS 0 o o o o o o o o o o o o o o o o o o o o o o o o o o 56 Introduction. . . . . . . . . . . . . . . . . . . . . . . 56 Questionnaire Returns . . . . . . . . . . . . . . . 58 Regression Equation to Predict the Number of Projects. . . . . . . . . . . . . . . . . . . . . . 59 Regression Equation to Predict a Combined Success Score. . . . . . . . . . . . . . . . . . . . . 63 Discriminant Function Analysis to Discriminate Between Innovators and Non-Innovators. . . . . . . . . 64 Discriminant Function Analysis of the Innovators Not Supported by EDP . . . . . . . . . . . . . . . . . 70 Classifying the Unsupported Innovators on the Basis of the Differences Between Innovators and Non-innovators. . . . . . . . . . . . . . . . . 70 Comparison with EDP-supported innovators . . . . . . . 72 Comparison with Non-innovators . . . . . . . . . . . . 77 Effects of Maturation . . . . . . . . . . . . . . . . . . 80 Major Findings. . . . . . . . . . . . . . . . . . . . . . 82 Summary . . . . . . . . . . . . . . . . . . . . . . . . . 82 SWY AND DISCUSSION 0 O O O O O O O O O O O O O O O O O O 86 Introduction . . . . . . . . . . . . . . . . . . . . 86 Conclusion No. l . . . . . . . . . . . . . . . . . 33 Implications of Conclusion No. 91 Conclusion No. 2 . . . . . . . . Implications of Conclusion No. Conclusion No. 3 . . . . . . . . . . . . . . . . . . . 93 Implications of Conclusion No. . . 99 Conclusion No. 4 . . . . . . . . . . . . . . . . . . . . 99 Implications of Conclusion No. 4 . . . . . . . . . . . 102 O O O O O O O O O O C 93 . . . . . . . . . . . 95 n:- F‘- O O O O we 0 iv Chapter Future Research . . . . . . . . . . . Research Including Other Variables Effect of EDP Support Over Time. . Replication with Better Controls . Summary . . . . . . . . . . . . . . . BIBLIOGRAPHY. . . . . . . . . . . . . . . . . . APPENDIX A. . . . . . . . . . . . . . . . . . . APPENDIX B. . . . . . . . . . . . . . . . . . . APPENDIX C. . . . . . . . . . . . . . . . . . . APPENDIX D. . . . . . . . . . . . . . . . . . . ”Pme E. O O O O O O O O O O O O O O O O O O Page 103 103 103 104 105 106 110 118 125 135 143 Table 10 ll 12 13 LIST OF TABLES MULTIPLE CORRELATIONS OF VARIABLES IN REGRESSION EQUATION TO PREDICT THE NUMBER OF PROJECTS . . . . . . . . UNSTANDARDIZED REGRESSION COEFFICIENTS TO PREDICT THE WER 0F PROJECTS 0 C I O O O O I O C . C O O O O O O UNIVARIATE F TESTS OF VARIABLE MEANS FOR INNOVATORS AND NON-INNOVATORS o o o o o o o o o o o o o o o o o o o o MULTIPLE CORRELATIONS AMONG VARIABLES IN DISCRIMINANT FUNCTION T0 DISCRIMINATE BETWEEN INNOVATORS AND NON- INNOVATORS o o o o o o o o o o o o o o o o o o o o o o DISCRIMINANT AND CLASSIFICATION COEFFICIENTS TO DISCRIMINATE BETWEEN INNOVATORS AND NON-INNOVATORS . . . . SIGNIFICANT UNIVARIATE F TESTS OF VARIABLE MEANS FOR EDP- INNOVATORS AND UNSUPPORTED INNOVATORS o o o o o o o o o SIGNIFICANT MULTIPLE CORRELATIONS AMONG VARIABLES IN THE DISCRIMINANT FUNCTION TO DISCRIMINATE BETWEEN EDP-INNOVATORS AND UNSUPPORTED INNOVATORS. . . . . . . . . DISCRIMINANT AND CLASSIFICATION COEFFICIENTS TO DISCRIMINANTE BETWEEN EDP-SUPPORTED INNOVATORS AND UNSIIPPORTED INNOVATORS O O O O O O O O O C C O I O O O O O SIGNIFICANT UNIVARIATE F TESTS OF VARIABLE MEANS FOR UNSUPPORTED INNOVATORS AND NON-INNOVATORS. . . . . . . . . SIGNIFICANT MULTIPLE CORRELATIONS AMONG VARIABLES IN THE DISCRIMINANT FUNCTION TO DISCRIMINATE BETWEEN UNSUPPORTED INNOVATORS AND NON-INNOVATORS. o o o o o o o o o o o o o o DISCRIMINANT AND CLASSIFICATION COEFFICIENTS TO DISCRIMINATE BETWEEN UNSUPPORTED INNOVATORS AND NON-INNOVATORS. . . . . COMPARISON OF CURRENT ACTIVITIES WITH FIVE PREVIOUSLY . . . . CHI SQUARE TEST OF PERCENT OF FACULTY REPORTING CHANGE IN CUMNT ACTIVITIES C C C C O I O O O O C C O O O O O O 0 vi Page 62 63 66 68 71 73 75 76 77 78 80 81. 81 LIST OF APPENDICES Appendix Page A QUESTIONNAIRE SENT TO EDP PROJECT DIRECTORS . . . . . . . . 110 B QUESTIONNAIRE SENT TO NON-EDP SUPPORTED FACULTY . . . . . . 118 C SUMMARY DATA FROM EDP-PROJECT DIRECTOR QUESTIONNAIRE RETURNS. C O O O O O O C O O O I O O O O O O O O O O O 125 D SUMMARY DATA FROM NON-INNOVATORS QUESTIONNAIRE RETURNS. . . 135 E SUMMARY DATA FROM UNSUPPORTED INNOVATORS QUESTIONNAIRE “TURNS O O O O O O O O O O O O O O O O O C 0 O O 0 O O 143 vii CHAPTER I THE PROBLEM Introduction A number of forces have been applying stress to institutions of higher education making it difficult to improve instruction (see Berquist & Phillips, 1975; Brown & Hanger, 1975; Dietrich & Johnson, 1967; Winstead, Draine, & Romine, 1969). Funds for improving instruction and faculty mobility have been declining, making it more difficult to introduce new ideas and techniques into colleges and universities. Demands for accountability and improved curriculum have been increasing. These forces have worked against the success of piecemeal efforts to improve instruction. In a recent article on faculty development, Berquist and Phillips (1975, p. 178) point out that a comprehensive approach to educational development is needed in higher education to deal with these problems. This comprehensive approach would involve the faculty, the instruction, and the organization simultaneously. The problem with implementing an innovative program of this type, however, is determining how to have the innovation accepted or adapted by the faculty. Rogers (1968, p. 71) identifies four types of decisions regarding the adoption of innovations which occur in a complex organization such as an institution or higher education. First, there are Optional decisions initiated by the individual regardless of the decisions made by colleagues. Second, there are contingent 1 decisions which require some prior decision to adopt or accent the innovation from the other members of the social system. Third, there are collective decisions which require consensus. Finally, there are authority decisions which are forced on the individual by someone with greater power. Rogers (1968, p. 71) goes on to suggest that one of the reasons that change is slower in higher education than in other fields is that the decisions to adopt innovations are contingent or collective rather than of the optional type. If a comprehensive approach to educational development (i.e., an innovation) is to be adopted on the basis of optional faculty decisions, it is likely that some faculty might not choose to adopt this innovation and would not participate in the educational deve10pment process. It is also likely that those who chose to participate probably differ on important dimensions from those who did not. Thus a research question of interest to those responsible for the diffusion and adoption of innovators in higher education is: Egg do innovators differ from non-innovators? In this study, innovators were defined as those who participated in an educational development program based primarily on optional innovation adoption decisions and non- innovators were those who did not participate. Purpose This study had two purposes. The first purpose was to identify in a higher education setting, specific characteristics of innovators and their departments which differed from non-innovators and their departments. The second purpose was to validate part of Rogers' and Shoemaker's model (1971) which deals with the characteristics of adapters of innovations. Validation is necessary since the contribution of data from higher education to the development of this model was extremely limited. Importance There are two reasons why identifying how innovators differ from non-innovators is important. First, information about the characteristics of the faculty reached by an educational development program based on optional innovation decisions can be used for the design of long range program evaluation and as a management tool to assist in providing more effective services. Second, it will identify which parts of the model on the adoption of innovations can be used to facilitate adoption of innovations in higher education. Research on the adoption of innovations process in education has provided incomplete support for the Rogers and Shoemaker model (1971). This problem can be traced to several causes. Much of this research failed to examine the influence of the adopter's social system on the adoption decision, thus overlooking, a large part of the model. Also, this research usually studied adoption by a school rather than adoption by an individual. Finally, most of the research was limited to public schools and did not explore the innovation process in higher education. This study addresses those weaknesses. Generalizability Even though this study was exploratory in nature, its findings should apply to other higher education settings other than the one studied. For example, because the research was conducted at a large university, faculty members may be more similar to colleagues in the same discipline at other universities than like colleagues in different disciplines at their own university. Haines (1974) found differences between disciplines greater than differences between institutions in a study on curriculum and instructional differences. Another reason that the findings of this study should be generalizable is that the research did not focus on the adoption of a specific innovation, such as computer assisted instruction. Therefore, the findings should be applicable to the adoption of innovations in general where adoption is based on an optional decision. Finally, the findings from this study can be used to support or refute the generalizations on the characteristics of adopters developed by Rogers and Shoemaker (1971) based on findings from research on the adoption of many innovators in many fields. Underlying Theory In their review of the literature, Rogers and Shoemaker (1971, p.61) identify three shortcomings of the existing research on the adoption of innovations in the field of education. 1. Failure to consider communication channels 2. Failure to consider how the social structure acts to impede or facilitate diffusion and adoption 3. Failure to recognize that adoption decisions are generally not optional decisions (i.e., they require approval or prior adoption by others in the social system) They also point out that the study of diffusion and ad0ption of innovations in the educational field is one of the weakest areas of existing research in terms of its contribution to the knowledge about diffusion and adoption. Traditional studies of the diffusion and adoption of innovations in education such as Brickell (1961) and Miles (1964) have focused mainly on individual variables such as socioeconomic factors, social participation, opinion leadership, etc. The flaw in these studies is that they fail to account for variables in the social system which cause the adopter under pressure to behave in ways not necessarily consistent with his personality. These pressure inducing social system variables include such things as the norms of the system, the decision making structure, style of leadership, etc. A further weakness of these traditional studies is that the link between the social system and the individual has not been explored. This link involves communication variables such as direction of communication, communication channel usage, adequacy of communication, etc. A number of researchers have recognized this flaw and begun providing models which take into account more of the elements in the adoption of innovations process (Carlson, 1968; Haveldck, 1969; Rogers & Jain, 1968; Sarbaugh & Hawkins, 1973). Yet, the greatest detail in identification of the elements in the innovation model is still provided by Rogers and Shoemaker (1971). Evidence of this is provided by the frequency with which Rogers and Shoemaker are cited as the underlying model for empirical and non-empirical work in this field. Havelock, a prolific writer in the field, goes into great detail; however, it is with reference to research utilization and the importance of the link between research and practice rather than with the diffusion and adoption of innovations or with the characteristics of innovators. By combining various descriptions of the innovation adoption process presented by Rogers (Rogers, 1968; Rogers & Jain, 1968; Rogers & Shoemaker, 1971), four interrelated categories of variables can be identified. 1. Individual variables 2. Social system variables 3. Communication variables 4. Consequences of adoption variables Within each of these categories are a number of variables Rogers and Shoemaker consider most significant in the innovation adoption process; definitions of those variables from the model which are of interest to this study are provided in a later section of this Chapter. The variables in each of the categories of this model are listed below: I. INDIVIDUAL VARIABLES l. Socioeconomic a. Age b. Education c. Literacy d. Status e. Upward social mobility f. Size of unit g. Commercial orientation h. Attitude toward credit 1. Specialization of operation Personality a. Empathy b. Dogmatism c. Ability to deal with abstraction d. Rationality e. Intelligence f. Attitude toward change g. Attitude toward risk h. Attitude toward education 1. Attitude toward science j. Fatalism k. Achievement motivation l. Aspirations m. Innovativeness Communication Behavior a. Social participation b. Integration with social system c. Cosmopoliteness d. Change agent contact. e. Exposure to interpersonal communication f. Exposure to interpersonal communication g. Information seeking about innovations h. Knowledge about innovations 1. Opinion leadership j. Membership in modern system k. Membership in integrated system II. SOCIAL SYSTEM VARIABLES 1. Norms 2. Leadership 3. Decision making structure 4. Characteristics of system that affect behavior III. COMMUNICATION VARIABLES 1. Amount of relevant communication in the system 2. Direction of communication 3. Symmetry of communication flow 4. Channels IV. CONSEQUENCES VARIABLES 1. Productivity or quality 2. Efficiency 3. Morale 4. Self-renewal Variable Selection The selection of variables from the Rogers and Shoemaker model for inclusion in this study was based on three criteria. First, the variable had to be an individual or social system variable; second, it had to be operationalized in such a way that its measure could be based on direct personal experiences or perceptions of the individual; and third, the variable had to be capable of being measured quickly. Although Rogers and Shoemaker identify four categories of variables only individual and social system variables were included in this study. It would have been inappropriate to include communica- tion variables since the unit of analysis in this study was the individual, and the unit of analysis for communication variables should be the dyad or social system. Consequences variables were similarily inappropriate since the unit of analysis for these variables should be the social system rather than the individual (Carlson, 1968; Rogers & Jain, 1968). The measurement of variables was limited to reports of direct personal experiences or perceptions by the respondents. The reason for this limitation was that validated measures did not exist for many of the abstract variables, and the instruments that did exist were often inappropriate for use in higher education settings. Development of these kinds of measures was beyond the scope of this study. As a third criteria, each variable had to be capable of being measured quickly. Since the emphasis of this study was on breadth rather than depth of analysis and since each faculty member would have a limited amount of time to deal with this study, keeping the measure short and the questions direct would encourage complete responses to more items by a large number of faculty. Along with these three criteria, two assumptions were made which affected the development of the research instrument used to measure these variables. First, it was assumed that the faculty member would have or be able to remember the information asked for, especially when asked about events within his department. Second, I it was assumed that faculty would be truthful and there would be noif systematic attempt to bias responses to obtain a future reward or 4/ to please the Educational Development Program which was sponsoring a part of this research. Applying these three criteria, and taking into account the 10 two assumptions regarding faculty responses, 18 of the 45 variables from the Rogers and Shoemaker model were selected for this study. These 18 variables were: age, status, size of teaching load, specialization of teaching responsibility, fatalism, innovativeness, social participation with departmental colleagues, intergration with their social system, cosmopoliteness, information seeking about instruction, knowledge about instructional innovations, opinion leadership, membership in modern and/or integrated system, norms on the importance of teaching, norms on innovativeness, norms on instructional strategies, resources for instructional improvement, and stability of instructional assignments. The Research Questions This study attempted to determine the answers to the following two research questions: 1. On which of the individual characteristics selected for this study do instructional innovators differ from non-innovators? 2. On which of the social system characteristics selected for this study do instructional innovators' perceptions of their departments differ from those of non-innovators? Innovators vs. Early Adopters Rogers and Shoemaker (1971, pp. 175—185) makes a distinction between innovators and other adopters of an innovation. They identify five categories, or ideal types, of adopters based on time of adoption: innovators, early adopters, early majority, late majority and laggards. According to this classification scheme, innovators have 11 different characteristics than early adopters. The primary differences are that the early adopter has more opinion leadership, is more in line with the norms of the social system, and tends to be less cosmopolite. However, a study by Stern, Craig, LaGreca & Salem (1976) at Ohio State University, failed to find these differences among pioneers, early, and late adopters of an innovation. Even Rogers and Shoemaker are not clear in the distinction between innovators and early adopters when they provide generalizations of research findings (Rogers & Shoemaker, 1971, pp. 352-376). They refer to "earlier" adopters rather than "early" adopters. In this study, MSU faculty who voluntarily conducted Educational Development Projects (the EDP project directors) were referred to as "innovators". The term innovator was considered comparable to Rogers' and Shoemaker's term "earlier adopters" for several reasons. First, Rogers and Shoemaker blur the distinction. Second, there was no way to measure the time of adoption to determine exactly how soon these individuals adopted an innovation after first becoming aware of it, so the group sampled undoubtedly contained both types. Third, this study did not compare different classes of adopters, but instead, it compared adopters with non-adopters. Since innovators would be more like early adopters than like non-adopters (Stern et a1, 1976) making a distinction between innovators and early adopters was not crucial to this study. However, the use of the label "innovators" for this group was more apprOpriate and descriptive than the label "early adopters". This is because there have been relatively few faculty supported by EDP, as compared to the total faculty at the University. This agrees 12 with the definition of innovators being among the first to adopt an innovation. Furthermore, the EDP criteria for selection and support of projects is biased toward those projects which are "innovative" within a_given department. The question of whether EDP project directors were innovators or early adopters is discussed in greater detail as a problem of selection in the section dealing with the research design in Chapter III. Definitions of Terms Fatalism. "Fatalism is the degree to which an individual perceives a lack of ability to control his future" (Rogers & Shoemaker, 1971, p. 188). Integration. "Communication integration is the degree to which the units in a social system are interconnected by interpersonal communication channels" (Rogers & Shoemaker, 1971, p. 188). Cosmopoliteness. Cosmopoliteness is the degree to which an individual's reference groups, or influences, are from outside the social system. (Gouldner, 1957, p. 290; Rogers & Shoemaker, 1971, p. 189). _Qpinion leadership. "Opinion leadership is the degree to which an individual is able to informally influence other individuals' attitudes or overt behavior in a desired way with relative frequency" (Rogers & Shoemaker, 1971, p. 35). Modern systems. "A modern system is typified by: l. A generally positive attitude toward change. 2. A well developed technology with a complex division of labor. l3 3. A high value on education and science. 4. Rational and businesslike social relationships rather than emotional and affective. 5. Cosmopolite perspectives, in that members of the system often interact with outsiders, facilitating the entrance of new ideas into the social system. 6. Empathic ability on the part of the system's members, who are able to see themselves in roles quite different than their own" (Rogers & Shoemaker, 1971, pp. 32—33). Overview of the Study The review of related research in Chapter II compares the findings of empirical studies on the diffusion and adoption of instruc- tional innovations in higher education with the findings from a variety of fields and draws implications of this research for the present study. The methods used to collect and analyze the data from this study are presented in Chapter III. An explanation and rationale for the multi-variate analysis techniques which were used is also provided. The findings and appropriate discussion are presented in the final two chapters. Chapter IV describes the results of the various multi-variate analyses performed. Chapter V discusses these results and their implications, compares the findings of this study with the findings from the related research presented in Chapter II, draws conclusions and makes recommendations for further research. CHAPTER II REVIEW OF RELATED LITERATURE Introduction This review of the literature will present research dealing with the characteristics of adopters and non-adopters of instructional innovations in higher education. It will also compare the findings from higher educations with generalizations from similar research done in many fields. As pointed Out in Chapter I, there is not a clear distinction between the terms "innovator" and "adopter". Since most of the literature uses the term "adopter", that term will be used in this chapter. However, the term "innovator" is still most descriptive of the adapters involved in this study. The review will show that there is a very limited amount of research literature pertaining specifically to the diffussion and adoption of innovations in higher education and even less that deals with the characteristics of adopters and non-adopters of innovations. Furthermore, the research that has been done often fails to deal with the actual adoption of innovations, dealing istead with attitudes toward the specific innovation. In general, the findings from studies in higher education do not provide much support for Rogers' and Shoemakers' (1971) generalizations about the diffusion and adoption of innovations. l4 15 Overview of the Literature The literature dealing with the diffusion and adoption of educational innovations is limited in scope and primarily focused on public school systems. The most extensive review of empirical studies is found in Rogers and Shoemaker (1971). They draw on approximately 1,200 empirical studies to derive 103 generalizations about the process of the diffusion and adoption of innovations. Approximately 1,084 of those studies were available in the Diffusion Documents Center (then at Michigan State University). Rogers and Shoemaker (1971, p.50) indicate that only 71 of these studies dealt with educational innovation, most of which center on the innovation process in the public schools. Carlson (1968), in addition to Rogers and Shoemaker (1971), points out a number of problems with the existing literature on the diffusion and adoption of innovations. First, many of the studies dealing with educational innovation have lacked rigor making it difficult to rely on the findings that do exist. Carlson goes on to indicate that no one study has examined all the elements of the diffusion and adoption process. Among the elements often omitted in the area of educational innovation research are channels of communication, social structure, and the value systems involved (Carlson, 1968, p. 5). Another problem with the research is that most of it has focused on the local school system rather than on the individual teacher as the unit of adoption (p. 14). Finally, there has been little concern for whether the adoption decision involved in the study was of the optional, contingent, collective, or authority type. Rogers (1968, p. 70) suggests that the type of decision greatly affects the rate of adoption in the social system. 16 This review will be limited to literature dealing with four- year institutions of higher education since these institutions represent social systems having important differences from the public schools and community colleges on several of the variables of interest in this study. Variables such as size of teaching load, specialization of teaching responsibility, and norms on the importance of teaching may differ among these social systems since faculty responsibilities in terms of teaching, advising, research, and service differ. These factors play an important role in determining the impact of the social system on the diffusion and adoption process (see Carlson, 1968; Rogers, 1968; Rogers & Shoemaker, 1971). Therefore, limiting the review to literature dealing with four-year institutions of higher education will allow the unique effects of the social system on the diffusion and adoption process to be identified. This review of the literature will also be limited to empirical studies so that the contributions of research on the diffusion and adoption of innovations in higher education can be compared to generalizations drawn from similar research in other social systems (see Rogers & Shoemaker, 1971). Among the problems encountered in reviewing the literature to identify specific findings from higher education is that the major references in the field (Brickell, 1961; Miles, 1964; Rogers, 1962; Rogers & Shoemaker, 1971) either fail to provide empirical data which can be used to draw generalizations or fail to identify the findings unique to higher education. Brickell limits discussion of higher education to its influence on the diffusion and adoption process in the public school 17 system. Miles put together a series of 25 papers from a seminar involv- ing the faculty of Columbia University. Ten of these were case studies of particular innovations and provided little evidence from which to draw meaningful generalizations applicable to the study of innovators in higher education. An additional nine papers dealt with research and theory, reporting on studies in public school systems or on evaluation of the effectiveness of particular innovations. The remaining six papers dealt with the American educational system and fail to shed any direct light on the process of the diffusion and adaption of innovations in higher education. Rogers, both in his 1962 book and his 1971 book with Shoemaker, fails to separate the findings from either education or higher education from the findings from other disciplines and cultures in the process of deriving generalizations from empirical data. Organization of the Review This review will be divided into four sections. The first section will describe research on characteristics of adopters and non-adopters of innovations. The second section will examine studies dealing with attitudes toward innovation rather than the actual adoption of innovations. The third section will present several studies related to the variables of this study, but which did not differentiate between adopters and non-adopters of innovations. The final section will compare this review with the generalizations Rogers and Shoemaker have drawn from the research of many fields. 18 Studies of Adopters and Non-adopters It was possible to locate reports on only two studies which compared actual adopters of an innovation with non-adopters. A study of ten departments in science and engineering at Ohio State University (Stern et a1., 1976) studied the characteristics of users of a new mechanized information center who were compared with a group of non-users (adopters and non-adopters) in terms of cosmopoliteness, attitudes toward change, professional background, scholarly productivity, age, rank, years with degree of years at Ohio State. However, they did find that adOpters sought more reinforcement from their peers. This was the case in both the adopter's professional and friendship networks. They also failed to find any significant differences between different classes of adopters based on time of adoption (pioneer, early, late). Mitchell (1970) used a different unit of analysis in a study of curriculum innovation at Northern Michigan University. He examined the social system before and after the introduction, adoption, and partial discontinuance of the innovation. He also found that as adoption or acceptance of the innovation increased, integration of the system as indicated by the communication network also increased. He also found that the lack of knowledge and lack of adequate resources and support within the administration contributed to negative attitudes and ultimate discontinuation. This is summed up in one of his conclud- ing generalizations, "An innovation is more likely to be adopted into {/ a complex organizational system, and is more likely to be continued / in use after adoption, if it is compatible with the existing norms and operational practices of the system" (p. 222). The adoption l9 decisions for this innovation, however, were not of the optional type and his findings provide little information about the individual adopter. Studies of Attitudes Toward Innovation There were five reports on studies dealing with attitudes or receptivity to adoption rather than with adoption itself. In the Stern et a1. study (1976), though, they found that actual adoption of an innovation differed from statements of intention to adopt. In one of the most complete reports, Kazlow and Giacquinta (1974) examined the receptivity of the faculty in a School of Education to a variety of organizational innovations. They used a variety of status and personality measures to explain differences in receptivity. While measures of status accounted for more variance than measures of personality, most of the variance remained unexplained. They also found that receptivity was innovation specific, dependent on the perceived risks associated with adoption. Of all their variables, I only differences in academic rank and sex were statistically significant for all innovations, though direction of these differences cannot be determined from the report of their study. Other variables used in their study included administrative rank, level of instruction, tenure, advisement load, research, teaching preferences, number of publications in the past five years, faculty group affiliation (cosmopolitan or local), dogmatism, attitude toward change in work, and related activities. A major limitation in generalizing from this study is that none of the innovations were of the optional decision type. They included such things as student involvement in academic governance, 20 reorganization of the school into a graduate school, establishment of a university without walls, undergraduate program, etc. Evans and Leppmann (1968), in a frequently cited book, report on a study on attitudes toward instructional television (ITV) at a State University. They compared the characteristics of pro-ITV faculty with those of anti-ITV faculty. Pro-ITV faculty were found to be more flexible in their attitudes and more concerned with teaching methods. Anti-ITV faculty were found to hold more traditional attitudes, be less willing to support change, and carried larger teaching loads. They also found that those who were generally receptive to new ideas from outside the university were more receptive to ITV. Resistance to ITV appeared greatest in the humanities rather than in technological fields. The disturbing thing about this study is that data in support of most of these findings is not presented in a manner that allows careful scrutiny. It seems that most of the conclusions are attempts to interpret and explain the data. The authors even point out they are describing ideal types rather than actual faculty members. Demerath and Daniels (1973) in a review of the literature and series of interviews found differences between academic disciplines in terms of receptivity to and interest in electronic innovations. The physical sciences showed most interest while the social sciences and humanities showed the least. It was hypothesized that this difference occurred because of the physical scientists' familiarity with technology as compared with the social sentiment that technology has ruined the environment. While their interpretation may be overstated, they appear to have found some effect due to the differences 21 between the norms of the various disciplines. Two studies identified particular attitudes that acted to unpede instructional innovation in higher education. These attitudes represent fatalistic beliefs on the part of individual faculty members (see Rogers & Svenning, 1969, p. 280). Neither is reported in any detail and both show very similar results. Cleland (1969) identified seven.attitudes that impeded the development of a college level program to train new college and university teachers. The one dealing with instruction was, "College teaching is an art that cannot really be taught, but is something that develops through long practice" (p. 424). Mangano (1973) reports on a study by Lora Robinson that identified three other attitudes relevant to instructional innovation, "Good teachers are born, not made"; "Teaching is an art, not a science"; and "Teaching is something you do, not something you talk about" (p. 208). Other Studies There are a number of other studies that deal with the variables involved in the diffusion-adoption process, but which do not make distinctions between adopters and non-adopters. Rogers (1968, p. 71) cites a study at Michigan State University which found faculty members were most knowledgeable about innovations that directly affected them. However, only one of the five innovations was directly related to instruction and the differences in knowledge levels was not made clear. Davis (1965) examined differences between the faculty at Private, liberal arts colleges which adopted many innovations (innova- tive colleges) and those which had not (non-innovative colleges). He fOund no significant differences in terms of awareness of the innova— tions or in terms of such individual variables as age, years of service, 22 and participation in decision making. There were differences, however, between the norms in the two types of colleges as the more innovative colleges were more permissive toward faculty members. It was unclear, though, whether there was any relationship between being on the faculty of an innovative college and being an innovator. Gouldner (1957) was involved in a slightly different type of study. He was not interested in innovation. Instead, he was studying cosmopoliteness. He identified three characteristics which could be used to differentiate cosmopolitans from locals: loyalty to employing organization, committment to specialized or professional skills, and reference group orientations (p. 290). Most other definitions of cosmopoliteness deal only with reference group orientations (Rogers & Shoemaker, 1971). In a study at a small, private liberal arts college, Gouldner found support for this distinction. Cosmopolitans were low on loyalty to the employing organization, high on committment to specialized skills, and likely to use outside reference group orienta- tion. Cosmopolitans in his study were more likely than locals to get most of their intellectual stimulation from sources outside the college. He also found that extreme locals tended to participate in the system more than extreme cosmopolitans, yet those in the middle participated most of all (p. 298). James (1975) conducted a study involving student evaluations of 268 faculty members in a College of Arts and Sciences. The evalua- tions were provided by 7,901 students. There were significant differences between tenured and non-tenured faculty (non-tenured receiving better ratings) and between academic ranks (associate professors receiving the highest ratings). Even though there was no 23 direct evidence linking these findings to innovation, they point out potential differences related to status which may have been due to innovation (since improved teaching could be a goal of innovation). The final group of three studies to be dealt with in this review identify differences between academic disciplines (Haines, 1974; Lewis, 1967; Peters, 1972). All three studies dealt with faculty perceptions of their curriculum. Haines found differences in terms of the perceived importance of different types of uses of cognitive information when the faculty were divided into three disciplines: humanities, social sciences, and natural sciences. Lewis added a forth discipline - engineering - and looked at attitudes toward what the university should offer and the perceived importance of various faculty activities (teaching undergraduates, teaching graduates, and research). He was able to find consistent differences between the humanities and engineering in terms of what the university should offer. However, differences between the various activities were not well supported and the findings in terms of the disciplines science and social science were very inconsistent. Peters used a more complex design and found faculty from engineering, agriculture, chemistry, and math identified with a more structured curriculum while faculty from english, speech, education, and psychology identified with a less structured curriculum. History, government, business management, accounting, and economics fell between the two extremes. The findings that show differences between disciplines may be due to inherent differences in the subject matter or due to differences between the social systems. There has been a significant amount of research on diffusion and adoption of ideas within the social systems of disciplines (see Crane, 1972) which 24 indicates that there are communication relationships within academic disciplines composed of departments at different universities that go beyond the relationships that exist among departments of the same university, e.g., the utilization of previous work (often an innovation) is governed by the norms of the "invisible college" (a communication network linking members of an academic discipline) rather than by the norms of an academic department at a single university (p. 83). Comparison with Findings from Other Fields The research findings relating diffusion of innovations in a variety of other fields are summarized in the generalizations presented by Rogers and Shoemaker (1971, pp. 347-385). For each of the variables included in this study, the appropriate generalization will be presented along with an indication of the support for it which comes from empirical studies in higher education. The variables will be grouped in the four categories Rogers and Shoemaker use to classify the individ- ual and social system variables of interest to this study: -socioeconomic variables, personality variables, communication behavior variables, and social system variables. Socioeconomic Variables. Four variables are included in this category: age, status, size of teaching load, and specialization of teaching responsibility. The generalizations of research findings indicate the following relationships for adopters and non-adopters (or later adopters in Rogers and Shoemaker terminology). "Earlier adopters are no different than later adopters in age." "Earlier adopters have higher social status than later adopters." "Earlier adopters have larger sized units (farms, and so on) than do later adopters." 25 "Earlier adopters have more specialized operations than later adopters." Five of the studies from higher education deal with socio- economic variables (Davis, 1965; Evans & Leppmann, 1967; James, 1975; Kazlow & Giacquinta, 1974; Stern et al., 1976). The two studies deal- ing with age found no significant difference between adopters and non- adopters. Three studies dealt with an optional decision finding no significant differences based on status measured by academic rank, while the two other studies found significant differences based on rank or tenure. One study found the faculty with negative attitudes toward an innovation had a larger teaching load. Finally, one study looked at level of instruction (undergraduate or graduate instruction) and teaching preferences and found no significant differences between faculty in terms of receptivity to innovations. With the exception of age, there is scant support from higher education for the generalizations concerning socioeconomic variables, and there appears to be evidence contrary to the generalization that earlier adopters have larger sized units. Personality Variables. Two variables are included in this category: fatalism and innovativeness. The generalization for fatalism is, "Earlier adopters are less fatalistic than later adopters." Two studies from higher education (Cleland, 1969; Mangano, 1973) identify specific fatalistic attitudes which have acted to impede innovation. This lends support to the generalization concerning fatalism. There is no generalization for innovativeness because innovativeness is "the degree to which an individual is relatively earlier 26 in adopting new ideas than other members of his social system" (Rogers & Shoemaker, 1971, p. 175). The generalizations deal with the characteristics of early adopters (more innovative) as compared to later adopters (less innovative). No studies were located which looked specifically at differences in overall innovativeness among adopters and non-adopters or early and late adopters. Rogers (1962, p. 188) points out that it may be valuable to have an individual's self-rating of innovativeness since the individual may be acting in accord with that perception of himself. Communication Behavior. Seven variables are included in this category: social participation with departmental colleagues, integration with their social system, cosmopoliteness, information seeking about instruction, knowledge about instructional innovations, opinion leadership, and membership in modern or integrated systems. The generalizations from research findings indicate the following relation- ships for adopters and non-adopters. "Earlier adopters have more social participation than later adopters." "Earlier adopters are more highly integrated with the social system than later adopters." "Earlier adopters are more cosmopolite than later adopters." "Earlier adopters seek information about innovations more than later adopters." "Earlier adopters have greater knowledge of innovations than later adopters." "Earlier adopters have a higher degree of opinion leadership than later adopters." 27 "Earlier adopters are more likely to belong to systems with modern rather than traditional norms than are later adopters." "Earlier adopters are more likely to belong to well integrated systems than are later adopters." Six of the studies from higher education deal directly with communication behavior variables (Davis, 1965; Demerath & Daniels, 1973; Evans & Leppmann, 1968; Gouldner, 1957; Kazlow & Giacquinta, 1974; Stern et al., 1976). One study looked at part of the social participa- tion dimension, participation in decision making, and found no signifi- cant difference between faculty at innovative or non-innovative colleges. Three studies looked at the individual's integration with the social system. When the measure involved reinforcement from peers, there appeared to be a significant difference with adopters seeking more reinforcement; when the measure was years of service, there was not a significant difference. Three studies looked at cosmopoliteness, one found those with pro-adoption attitudes were more cosmOpolite while the other two found no significant differences. There were no studies that examined information seeking about instruction, knowledge about instructional innovations, or opinion leadership as they apply to adopters and non-adopters. One study did find, however, that faculty tend to be more knowledgeable about innovations that affect them. Three studies found differences between academic disciplines even though the studies were not directly related to innovation (Haines, 1974; Lewis, 1967; Peters, 1972).. These studies suggest there might be differences among discipline-oriented social systems. Only the generalization dealing with integration with the social system is 28 supported by the research from higher education, and that occurs only when the measure of integration involves reinforcement. Social System Variables. Two variables are included in this category: system norms and characteristics of the system. There is only one generalization that deals with social system variables, "System effects may be as important in explaining individual innovativeness as such individual characteris- tics as education, cosmopoliteness, and so on." Four studies deal specifically with system effects (Davis, 1965; Demerath & Daniels, 1973; Mitchell, 1970; Rogers, 1968). Two of the studies point out the effects of characteristics of the system which indicate differences among discipline-oriented social systems also support to this generalization (Haines, 1974; Lewis, 1967; Peters, 1972). This generalization appears to be well supported by the research from higher education. Summary This review of the literature has presented the research dealing with the characteristics of adopters and non-adopters of instructional innovations in higher education. It has also compared these findings from higher education with generalizations from similar research done in many fields. The review showed that there are very few studies dealing with the characteristics of adopters and non- adopters, and the research suffers from a lack of rigor and specificity with regard to the difference between attitudes toward an innovation and adoption of that innovation, as well as with regard to the type of adoption decision required by the innovation. It was also shown that none of the studies dealt with all the variables involved in the 29 diffusion and adoption process. Since there were so few studies from higher education, it is dangerous to generalize too far beyond them. However, they appear to support Rogers' and Shoemaker's (1971) generalizations concerning differences in age, fatalism, integration with the social system, effects between earlier adopters and later adopters. In addition, there is some evidence that the generalization concerning size of unit reversed for higher education. Implications for This Study There are six implications from the research for this study. First, the study should compare actual adopters or innovators with non-innovators since the actual adoption of an innovation may differ from statements of intention to adopt. Second, there is insufficient evidence from higher education to warrant discarding Rogers' and Shoemaker's model. Third, a study which looks at a large number of variables is called for since the diffusion and adoption process involves many variables which seem to be capable of differentiating between innovators and non-innovators. Fourth, the adoption decisions involved in the study should either be limited to a single type or else clearly separated since the type of decision affects the rate of adoption in the social system. Fifth, the study should involve faculty from a variety of disciplines rather than just one department, college or discipline since there appear to be differences among disciplines which could affect the diffusion and adoption process. Finally, the study should be conducted with rigor and reported with sufficient clarity to allow others to derive meaningful conclusions from it. CHAPTER III DESIGN OF THE STUDY Introduction To identify the differences between innovators and non-innova- tors, this study used an ex post facto multivariate comparison design (see Campbell 8 Stanley, 1963; Kerlinger & Pedhazur, 1973), utilizing data collected after the significant event has occurred (in this case, instructional innovation) to identify significant relationships between variables. It differed from true experimental research by its lack of controlled conditions. It is an appropriate design in this case, however, since it was not possible to control or manipulate all the factors or develop measures to establish a direct cause-effect relationship. The most obvious weakness of this design is that causes cannot truly be assigned and significant relationship may be masked by uncontrolled phenomenon. Furthermore, the analysis procedures used have the potential of identifying spurious relational patterns with little or no reliability or validity. None the less, the exploratory nature of this study warranted use of this design. The data were collected through questionnaires distributed to 345 faculty at Michigan State University. These faculty were divided into three groups for the purposes of analysis. The first group was composed of faculty who had received EDP grants for an instructional development project for improvement to an undergraduate course. Thes were the instructional innovators. The second group was composed of 30 31 faculty who reported that they had been involved in an instructional development project during the past five years, but who had not received EDP support. This group of unsupported innovators represented another group of innovators which could be used to confirm any differences found between the EDP-supported innovators and the non- innovators. The data were analyzed by two multiple regression techniques: simple multiple regression and discriminant function analysis. Multiple regression analysis is a method of multivariate data analysis which can be used to identify the effects of two or more independent variables, acting individually or in combination, on a dependent variable (see Kerlinger & Pedhazur, 1973, pp. 2-80; Kim & Kohout, 1975, pp. 320-367). While either analysis of variance of multiple regression analysis can be used when the independent variables are categorical, multiple regression analysis is the only appropriate method of analysis when: "1. The independent variable is continuous 2. The independent variables are both continuous and categorical 3. All frequencies are unequal and disproportionate 4. Studying trends in data: linear, quadratic, and so on" (Kerlinger & Pedhazur, 1973, p. 114). Since the variables in this study were both continuous and categorical, multiple regression was an appropriate method of analysis. Discriminant function analysis is a special form of multiple regression analysis. It is a method for distinguishing between two or more groups on the basis of discriminating variables. This is accomplished by forming one or more linear equations (functions) of the discriminating variables. These functions can then be used either for classification of individuals 32 to the groups or for analysis of the multivariate relationships between the variables (see Kerlinger & Pedhauzur, 1973, pp. 336-347; Klecka, 1975; Tatsuoka; 1971). This analysis occurred in four parts. First, a regression equation to predict the number of projects an innovator had done was generated from data supplied by the EDP project directors. Second, the same data that were used to generate the first regression equation were used to generate a second regression equation which was to produce a combined success measure for the innovator's first project. Both of these equations were then analyzed using discriminant function analysis to determine whether they could identify variables which would differentiate between innovators and non-innovators. The third part in the data analysis was the use of a separate discriminant function analysis to identify variables which differentiated between the known innovators (EDP project directors) and the known non-innovators. Finally, the data from faculty who had been involved in instructional development or instructional innovation projects without EDP support were compared with the data from the innovators and the non-innovators. This four-part analysis yielded information about which variables, acting individually or in combination, were significantly different between innovators and non-innovators, and also provided linear equations for use in future analysis of individual faculty members. Each variable that was useful in distinguishing between innovators and non-innovators was included in these linear equations, and their coefficients could be used to determine their relative importance in making the distinction, provided there are no problems with high intercorrelations among the variables. 33 The Research Design The design of this study may be illustrated as follows with O's representing the time period covered by the questionnaire used in this study, X's representing the time of instructional innovation, and T's representing the time of data collection. Innovators O X T Unsupported Innovators X OT Non-innovators OT Threats to internal validity may come from several areas. Since the instructional innovators will be recalling conditions in a different time period than the non-innovators, maturation effects may affect the two groups differently. An indication of the influence of maturity was provided by analysis of the responses to several questionnaire items comparing the individual's level of activity with that of five years previously. Large differences in activity levels would signify that maturity had a significant effect on any between group difference found. Selection may also pose problems for this study. The EDP project directors represented a self-selected sample since they choose to seek EDP support for instructional innovation through their own initiative. Furthermore, only a certain number were supported from those applying. EDP projects are selected for support on the basis of four general criteria: number of students affected, use of an experimental or innovative approach, generalizability, and capability for evaluation (Educational Development Program, 1975, pp. 7-8). This suggests that the EDP project directors may not be like other 34 instructional innovators such as those represented in the group of faculty in this study who reported participating in an instructional development project without EDP support. Analysis of the differences between these two groups of innovators will indicate the influence of selection on any differences found between innovators and non- innovators. Mortality of respondents may also threaten validity because t there may have been a differential loss of respondents among the ; groups. For example, the EDP—supported innovators may have felt a greater obligation to respond than those faculty not supported by EDP, and they may have biased their responses based on gratitude I, for the support of their innovation activities. In addition, a A number of instructional innovators had left the campus and may not I have had an equal opportunity or felt an equal obligation to respond as those still on campus. To the extent possible, identical procedures I were followed in the collection of data and subsequent follow-up to encourage participation with all groups. However, those who had left the university received no follow-up. Since there were a number of factors which could not be controlled in this study, the results should be viewed as exploratory rather than definative. Research Questions To answer the question raised in Chapter I of this study, "How do innovators differ from non-innovators in terms of 18 selected variables", this study will address two research questions: 1. On which of the following characteristics do instructional innovators differ from non-innovators? a. Age 35 b. Status c. Size of teaching load d. Specialization of teaching responsibility e. Fatalism f. Innovativeness g. Social participation with their social system h. Integration with their social system i. Cosmopoliteness j. Information seeking about instruction k. Knowledge about instructional innovations 1. Opinion leadership m. Membership in modern and/or integrated systems 2. On which of the following characteristics do instructional innovators' perceptions of their departments differ from those of non-innovators? a. Norms on the importance of teaching b. Norms on innovativeness c. Norms on instructional strategies d. Resources for instructional improvement e. Stability of instructional assignments Operational Definitions With only a few exceptions, as noted in the forthcoming section on item selection, there was little guidance available from past research. As a result, there were often several different ways to operationalize the variables of interest in this study. Each operationalization highlighted a different dimension of the variable. Therefore, for some of the variables in this study, more than one ‘1- —.._.‘ _- 36 operationalization or measure is provided. Each measure will be used separately in the analysis. However, in the event that these different measures are highly intercorrelated with one another (.8 to 1.0), only those not highly intercorrelated will be used in the final analysis (see Kim & Kohout, 1975, p. 341). The names and descriptions of the measures used in this study to operationalize the variables are as follows: ‘Agg, AGE was a self-report by respondents. Status. Two measures were used for this variable: RANK and TENURE. Each was a self-report by respondents. Size of teaching load. Two measures were used for this variable: COURSES and TEACHING TIME. Courses was a self-report of the number of different courses taught by the respondent. Teaching time was a self-report of the percentage of time devoted to teaching and indicated the size of teaching responsibility as measured by time as compared to other activities. Specialization of teaching_responsibility. Three measures were used for this variable: TEACHING ISOLATION, TEACHING SPECIALIZA— TION, and TEACHING IMPORTANCE. Teaching Isolation was a self-report on whether the respondent teaches unique or specialized courses-- courses not taught by other faculty members. Teaching Specialization was a combined measure of the number of courses taught and the uniqueness of those courses. A respondent is less specialized as the number of non-unique courses increases. Teaching importance was a self-rating of the relative importance of teaching compared to other activities. A respondent whose primary responsibility (speciality) ‘was teaching would rate teaching as an important activity compared to 37 other activities such as research. Fatalism. Two measures were used for this variable: TEACHER FATALISM and STUDENT FATALISM. Teacher Fatalism was measured by a self-rating of agreement with a fatalistic statement about good teachers. Student Fatalism was measured by a self-rating of agreement with a fatalistic statement about good students. The more fatalistic the individual, the greater the agreement with fatalistic statements. Innovativeness. INNOVATIVENESS was a self-rating on a scale from Rogers as described in the section on item selection. This scale provided an indication of when the respondent was most likely to adopt an innovation compared to colleagues. Social participation with departmental colleagges. PARTICIPA- TION was a self-report of the amount participation by the respondent in departmental meetings, seminars, social events, etc. Integration with the social system. Three measures were used for this variable: YEARS AT MSU, LOCAL INFORMATION, and LOCAL STYLE. Years at MSU was a self-report which represented the potential for forming the informal communication links necessary to be integrated with the social system. Local information was a self-report of the frequency of use for local sources of information (those from within the social system). A more integrated individual would make more use of local information sources. Local Style was a self-rating of the influence from various local sources on the respondent's teaching style. A more integrated individual would have greater local influence. Cosmopoliteness. Two measures were used for this variable: COSMOPOLITE INFORMATION and COSMOPOLITE STYLE. Cosmopolite Information was a self-report of the frequency of use for cosmopolite sources of 38 information (sources from outside this social system). A more cosmo- polite individual would make more use of cosmopolite information sources. Cosmopolite Style was a self-rating of the influence from various cosmopolite sources on the respondent's teaching style. A more cosmopolite individual would have greater cosmopolite influence. Information seeking about instruction. INFORMATION SEEKING was a measure of the frequency of use for both local and cosmopolite information sources. Knowledge about instructional innovations. KNOWLEDGE was a self-rating of the amount of knowledge about instructional innovations possessed by the respondent. Opinion leadership. Two measures, based on scales from Rogers described in the section on item selection, were used for this variable: OPINION LEADERSHIP and OPINION LEADERSHIP CHANGE. Opinion Leadership was a self-rating of the respondent's own credibility as a source of information. Opinion Leadership Change was a comparison of Opinion Leadership over time. Membership in modern and/or integrated systems. ACADEMIC DISCIPLINE was a classification of the respondent into one of three disciplines as described in the section on item selection. Norms on the importance of teaching. TEACHING VALUE was a rating of the respondent's perception of the importance or value of good teaching for promotion or other rewards within the department. Since norms tend to follow values, the more value placed on teaching, the stronger the possibility that the norms also support good teaching. Norms on innovativeness. FACULTY REACTION was a report on the reaction of faculty in the department to discussions about instructional 5/39 innovation. If the norms did not support innovation, the reaction would probably be negative. Norms on instructional strapggies. TEACHING MODELS was a report on the number of different teaching models used by colleagues in the department. If the norms supported a variety of teaching models, the number reported would be higher than if there were only a few "accepted" models. Resources for instructional improvement. RESOURCES was a rating of the perceived adequacy of resources available in the department for use in improving instruction. Stability of instruction asgdgpments. TEACHING STABILITY was a report on the frequency of changes in instructional assignments within the department. Population and Sample The population under investigation was comprised of faculty at Michigan State University. Two samples were selected. The first sample, the instructional innovators, included all EDP project directors identified in Reports 3-8 of the Educational Development Program at Michigan State. This included projects which took place between 1970 and 1975. The second sample was comprised of a random sample of the teaching faculty from throughout the University, which was later divided into two groups: unsupported innovators and non- innovators. Selection of the EDP project directors was based on the following criteria. First, the project director must have been listed in one of the six EDP reports as having received funds in support of a project. (All projects reported represented efforts which 40 had a direct influence on the instructional process in an undergradute course or curriculum.) Second, the EDP project director must have been a faculty member in an academic program whose primary responsibil- ity Egg pg£_instructional development. Third, the EDP project director must have been an assistant, associate, or full professor in the tenure stream at the time of the project. All EDP project directors who did not meet these three criteria were eliminated from the study, e.g., those involved in faculty or organizational development, or those who were only of instructor rank. This sample was composed of 90 faculty members. The random sample of teaching faculty was selected from the latest faculty roster of the university. Any faculty members who had been EDP project directors at any time or whose primary responsibility was instructional development were elimdnated. Selection of these faculty was based on the folloging criteria. Each faculty member selected must have been an assistant, associate, or full professor in the tenure stream at the university. They must not have held administrative responsibility above that of department chairman. They must have been at the university for at least two years. Finally, they must have taught at least two courses, at least one of which was an undergraduate course, during the academic year as identified in the official schedule of courses for 1975-76. A random sample of 250 faculty members was selected to ensure at least 10 respondents for each variable, based on a return of 70-752 of the questionnaires (see Allen, no date). 41 Instrumentation Questionnaire develppment and distribution. Two questionnaires were developed based on the operational definitions of the variables of interest in this study. Questions on the instrument sent to the EDP project directors asked that their responses be based on the time prior to their first EDP project, even if in the case of multiple projects the first project occurred before 1970. The instrument also contained items dealing with information outside the realm of this study which was for use by the EDP staff (see Davis, Abedor & Witt, 1976). The questionnaire was field tested with several faculty who were familiar with actual EDP projects but were not themselves EDP project directors. Revisions were made before distributing this instrument to faculty. This questionnaire appears in Appendix A. The second questionnaire was distributed to faculty not supported by EDP. The variables of interest in this study were measured using items identical to those on the first questionnaire with the exception that respondents were asked to base their answers on the preceding year. In addition, there was an item asking if the individual had been involved in an instructional development or instructional innovation project during the past five years, and several items asking for a comparison of their current levels of activity with their levels of activity five years previously. The items on past levels of activity were for use in assessing the impact of maturation. The item on participation in an instructional development project was used to identify non-innovators. There were also several items on teaching style which were not of importance to this study. This questionnaire appears in Appendix B. 42 The questionnaires were distributed by mail one month apart and were both accompanied by a letter from an Assistant Provost at the University who urged the faculty to participate in the survey. Questionnaires were sent to EDP project directors in March, the others in April. Telephone follow-ups were made to each faculty member who had not responded after ten days. If the faculty member could not be reached, a note was left with their secretary asking the faculty member to response. A second telephone follow-up was made a week later to all those faculty who had still not responded and who had not been spoken to directly in the first follow-up. Identical procedures were used with both questionnaires. Item selection. A number of questionnaire items were specifically selected based on previous research. Despite the findings of no signif- icant difference found for the variables age and years at the institution, these items were included in this study for comparison with other studies (Stern et al., 1976; Kazlow & Giacquinta, 1974). The item dealing with the teaching models used in the respondent's department was based on the list of teaching models used by the Michigan State University Office of Institutional Research. The items dealing with innovativeness and opinion leadership were taken from self-rating scales used by Rogers (1962, p. 188; Rogers & Shoemaker, 1971, pp. 215-217; Rogers & Svenning, 1969, p. 224). The two items dealing with fatalism were also adapted from a Rogers' scale (Rogers & Svenning, 1969, p. 280) using the find- ings from Cleland (1969) and Robinson (cited in Mongano, 1973) to provide the fatalistic statement about teachers. It was not possible to measure the degree of integration or modernness in the various social systems of which the participants were 43 a part. These measures would have required information from faculty not in the sample to provide an accurate assessment. Previous research suggested that there are systematic differences among academic disciplines (Haines, 1974; Lewis, 1967; Peters, 1972). None of these studies, however, dealt specifically with the diffusion and adoption of innovations. While there was no reason to believe the causes of these differences were either integration or nodernness, both explanations are possible. For the purposes of this study, three academic disciplines were used to represent social systems which could vary in terms of integration and modernness: natural science, social science and humanities. If innovators and non-innovators differed in terms of academic disciplines, it would indicate that some factor(s) in the nature of the social system were important (such as integration or modernness). If a significant contribution was not found, then the findings from the other variables dealing with the social system must be examined to judge whether the selection of social systems was inappropriate or whether social system factors, such as integration or modernness, were not significant. The distribution of colleges to the academic disciplines was as follows: Natural Science Agriculture and Natural Resources Engineering Human Medicine Lyman Briggs (a residential natural science college) Natural Science Veterinary Medicine 44 Social Science Business Communication Arts and Sciences Education James Madison (a residential social science college) Social Science Urban Development Humanities Arts and Letters Human Ecology Justin Morrill (a residential liberal arts college) University College (the undergraduate basis education college) The other items used on the questionnaires asked for reports or ratings of conditions in the respondent's department. Reliability and Validity It was not possible to conduct a test of reliability of the instrument used in this study. Such reliability tests as test- retest and equivalent form, could not be used since access to the respondents was limited to one administration of the questionnaire. Such tests as split-half or inter-item consistency were inappropriate since the instrument used in this study did not attempt to measure a single trait and the number of items which could be included on the instrument was limited (see Anastisi, 1954; pp. 94-117; Kerlinger, 1964, pp. 429-443). The validity of the instrument was measured by the ability of the various discriminant function analyses to correctly classify 45 innovators and non-innovators. Anastasi refers to this as "empirical validity" (1954, p. 127). This type of validity compares the individual's score (in this case, the regression or discriminant score) with some criterion which is a direct measure of the characteristic the instrument is attempting to predict (in this case, group membership). Data Analysis Overview. The data was coded and keypunched, and then analyzed using the MSU CDC 6500 computer and the Statistical Package for the Social Sciences (SPSS) Multiple Regression and Discriminant Function Analysis programs. This analysis was to determine whether innovators and non-innovators differed, on which variables they differed, and to identify combinations of variables that maximally separated the EDP project directors from the other faculty members. The following assumptions were made regarding the data in this analysis: 1. The data followed a multivariate normal distribution 2. There were equal variance-covariance matrices 3. Each faculty member provided an independent response 4. There was no systematic loss of respondents Klecka (1975, p. 435) points out, however, that the discriminant function analysis technique is very robust and the assumptions pertain- ing to the distribution and variance-covariance matrices do not have to be strictly adhered to. Development of multiple regression equations. Multiple regress- ion analysis provides the best linear equation to explain the data. It controls for confounding factors so that the contribution of a 46 variable or set of variables can be isolated, and it provides information on relationships among the variables. In line with previous research, a linear model was assumed for this study. This assumption was based on a discussion by Rogers and Svenning (1969, p. 57) about the model for this study. They point out, "There is little reason to anticipate curvilinear relations among variables; and in fact, such relationships are not encountered when the variables are checked empirically." The data from the EDP project directors were used to generate two multiple regression equations. One attempted to predict the number of EDP-supported projects conducted by an individual because if the innovators who did one project were more like the non-innovators than like the innovators who did many projects, then the regression equation might be composed of those variables on which innovators and non-innovators differed. The second regression equation attempted to predict a combined measure of project success because if less successful innovators were more like non-innovators than like more successful innovators, then the regression equation might be composed of variables, other than the ones used to compute the combined success score, on which innovators and non-innovators differed. The combined measure or project success was based on whether the results of the innovation were still in use; whether the EDP project director considered the project worth the effort, would be willing to do another project, tried other innovations since the project, saw increases in student learning and attitudes, and along with his colleagues, considered the Project a success. 47 Each of the two regression equations could be used to compute a regression score for each respondent. The regression score was computed by using each individual's responses as the values for the variables in the regression equation. To determine if innovators and non-innovators actually differed on the variables in the two regression equations a discriminant function analysis of the regression scores produced by each regression equation was used. This analysis procedure attempts to classify the respondents into two groups--in this case, innovators and non-innovators-- by comparing the individual's regression score with the regression scores of other individuals in the two groups. The ability of this procedure to separate the known innovators, the EDP project directors, from the known non-innovators provided a test of whether the variables in each regression equation represented variables on which innovators differed from non-innovators. Direct comparison usingrdiscriminant function analysis. Data from the EDP project directors and the non-innovators were also compared directly, without using an intermediate regression equation to identify variables which differed between the two groups. Instead of using a regression score based on an equation generated to predict some characteristic of innovators, such as number of projects done or a combined measure of success, the direct use of discriminant function analysis involved generating a regression equation (function) which predicted Which group an individual belonged to. Those variables which are included in this function (the discriminant function) are those variables which differentiate between the two groups: innovators and non-innovators. If this procedure could identify the unsupported I innova findir which advant Since from t the va be est: those 1 or inst functic ‘mnoval either not p1 discrt a tes grout ident data 5 1final c 66811 111% 48 innovators as belonging to a single group, empirical validity of its findings would be established. Using the two regression equations to identify the variables which differed between innovators and non-innovators has a major advantage over the direct use of discriminant function analysis. Since the regression equations were generated using only the data from the EDP project directors, their validity and the validity of the variables to predict between innovators and non-innovators can be established by classification of a known group of non-innovators, those who reported not being involved in an instructional development or instructional innovation project. However, since using discriminant function analysis directly requires that both innovators and non- innovators be used to generate the discriminant function, reclassifying either of the groups on the basis of this discriminant function does not provide an unbiased test of its predictive validity. Using the (discriminant function to classify the unsupported innovators does provide as test, though there is no way to determine in advance whether this Egroup should be classified as innovators, non-innovators, or neither. ESummary This study used an ex post facto multivariate design to ixientify the differences between innovators and non-innovators. The detta.were collected by questionnaires distributed to: EDP project disrectors from the past six years who have been involved in instruc- tional development projects for the improvement of an undergraduate conirse, and a randomly selected sample of 250 faculty who had not been involved with EDP supported projects. 49 These 250 faculty were later separated into two groups based on their responses to an item concerning their involvement in instructional innovation. Those that reported no involvement in instructional development or instructional innovation were classified as non-innovators. Those that reported that they had been involved in instructional development or instructional innovation without EDP support were used to test the predictive validity of the findings and provide an indication of the bias due to selection. All of the non-EDP supported faculty were asked several questions to identify the possible impact of maturation on the findings. The data were analyzed by several multiple regression techniques: simple multiple regression and discriminant function analysis. These procedures yielded information about which variables were significantly different between the two groups. The following two sections describe the specific procedures used in the multiple regression and discriminant function analyses performed in this study. The reader may skip these sections and proceed directly to Chapter IV without any loss in continuity. Multiple Regression Analysis This discussion of the procedures used to conduct the multiple regression analysis will describe four aspects of the analysis. First, the appropriate tests of significance will be described. Second, the criteria and process for including variables in the regression equation will be presented. Third, the procedures used to deal with missing data will be described. Finally, the relationship between the variance explained in the sample and the variance explained in the population will be indicated. 50 Significance tests. Two tests of significance were used for the multiple regression equations (Kim & Kohout, 1975). The first tested the goodness of fit of the linear equation. When this test is statistically significant, it indicates that the linear relationship between the variables as expressed in the regression equation is not due simply to chance (with a probability of error indicated by the level significance). This involved conducting an analysis of variance for R2, the amount of variance explained by the regression equation. A univariate F-test was used to test the null hypothesis that the multiple correlation observed was due strictly to chance; that is, Rz-O. The appropriate formula for this F-test is: Rz/k (14:2) / (N-k—l) a: ll k = number of independent variables, N = sample size, degress of freedom = k and N-k-l (p. 335) The second test of significance involved testing the specific regression coefficients. Only those variables whose coefficients were statistically significant were included in the equation. This required that R2 be broken down into components attributable to each independent variable in the equation. In this study, the standard regression method of decomposition was used. With the standard regression method, each variable was treated as if it was added in a separate step after the preceding variables were controlled for. Thus, the incremental increase in R2 is the component of R2 attributable to that variable. The appropriate test of significance for this regression coefficient is: 51 F = Incremental increase in R2 due to variable X (1-R2) 1 (N-k—l) k = number of independent variables, N = sample size, degrees of freedom = l and N-k-l (p. 336) The method of breaking R2 into components is an accepted method when the relationship among the variables is assumed to be non-causal (p. 336). Inclusion of variables. An important consideration in building a regression equation is the order in which the variables are included in the equation. In this study, variables were included in single steps with the variable explaining the most variance included first. The variable which explained the greatest amount of variance in combina- tion with the first variable was included second. In other words, the variable that explained the greatest amount of variance unexplained by variables already in the equation was entered in each step. This resulted in a subset of the total number of variables that yielded an optimal prediction equation with as few terms as possible. Several conditions were used to control this process of inclusion of variables. The regression coefficient had to be significant at the .01 level and the proportion of each variable's variance unexplained by variables already in the equation had to be greater than .001. In addition, it was possible for all variables to be entered in the equation at any step. This ensured that the maximum number of variables would be included (Kim & Kohout, 1975, p. 345). Replacement of missing_data. Only data from EDP project directors who returned complete responses for all the variables of interest in this study were used to generate the regression equations. This insured that partial correlations were computed from the same population. After the regression equations were generated, however, 52 regression scores were calculated for all respondents. At this time, missing data were replaced by the mean values of the missing variables. Population variance. The amount of variance explained by regression equation, R2, is an overestimate of the population R2. It is affected by a number of factors, among which is the ratio of the number of independent variables to the size of the sample. The formula to estimate the squared multiple correlation in the population from the sample multiple correlation is: (R2 (corrected for shrinkage) = l-(l-RZ) (NgRil) k = number of independent variables, N - size of both samples, R2 = sample R2 (Kerlinger & Pedhazur, 1973, p. 283). Discriminant Function Analysis This discussion of the procedures used to conduct the discriminant function analyses will describe six aspects of the process. First, the method for determining the number of discriminant functions necessary to differentiate between the groups and the nature of the coefficients in the discriminant functions will be described. Second, the process used to classify respondents to groups based on the variables in the regression equations and discriminant functions will be presented. Third, the appropriate tests of significance will be indicated. Fourth, the criteria and process for including variables in the discriminant functions will be described. Fifth, the procedures used to deal with missing data will be presented. Finally, the relationship between the variance explained in the sample and the variance explained in the population will be indicated. 53 Discriminant functions and coefficients. The maximum number of discriminating functions used to discriminate between groups is equal to one less than the number of groups, so long as there are more independent variables than groups (Klecka, 1975, p. 435). In this study, then, there was one discriminating function generated since there were two groups involved at each stage of the analysis. The variables used in the discriminating function were standardized, which means that their coefficients in the function are also standardized. This allows these coefficients to be compared to each other since each represents the relative contribution of the variable to the function. The classification_process. The discriminant function(s) are used to calculate discriminant scores by using each individual's responses for the values of the variables in the functions. These scores are necessary to produce the new functions actually used to classify respondents to the groups. Two new regression equations were generated to predict these discriminant scores. These new equations were the classification functions and their coefficients are the classification coefficients. A separate classification function was necessary for each group. Each classification function was used to compute a score for each respondent. The respondent was then assigned to the group whose classification function yielded the highest score for that individual. (An additional adjustment was made in the classification functions to account for the prior distribution of respondents to the two groups.) Significance tests. The appropriate statistic for discriminant function analysis is Wilks lambda. Wilks lambda is the relationship 54 between within group variance and total variance and is the most widely used test statistic in multivariate data analysis (Tatsuoka, 1971, p. 40). Wilks lambda is especially suited to cases in which there is more than one dependent variable, however, it can be used in cases with one dependent variable such as this study (Kerlinger & Pedhazur, 1973, p. 352). In multivariate cases, lambda is the equivalent of l-RZ. An appropriate test of lambda's significance is Chi Square (Tatsuoka, 1971, p. 430). The SPSS program provides both the Chi Square test for testing the significance of Wilks lambda considering the effect of the variables taken together, and a univariate F-test of the significance of the difference between the means of each variable taken separately. Inclusion of variables. As in the development of the multiple regression equations described in the previous section, variables were included in this discriminant function in stepwise fashion. The best discriminating variable was included first, followed by the variable which, combined with the first variable, explained the most variance, and so on. Several conditions were used to control this inclusion of variables. The test of discrimination used was the multivariate F-ratio of the difference between group centroids, which is the mean value of the discriminant function for each group. The minimum value of this Inultivariate F-ratio necessary for inclusion of a variable was 1.0. At the same time, the proportion of variance in a variable which was not explained by the variables already in the discriminant function had to be greater than .001. 55 Replacement of missing data. As in the multiple regression analysis, only respondents who provided complete data for all the variables were used to generate the discriminant function, though all the respondents were included in the classification computations-- missing data were replaced by the individual variable means. Population variance. Also, as in the multiple regression case, it is appropriate to use the formula already provided to estimate R2 for the population; however, (1-lambda) must be substituted for the value of sample R2. CHAPTER IV RESULTS Introduction This study set out to identify the differences between innovators and non-innovators in terms of 18 different variables. These variables involved 27 measures on questionnaires sent to EDP project directors and non-EDP supported faculty at Michigan State University. As indicated in Chapter III, the data from the questionnaires were analyzed in four parts. First, a regression equation to predict the number of projects an innovator had done was generated from data supplied by the EDP project directors. Second, the same data that were used to generate the first regression equation were used to generate a second regression equation which was to predict a combined success measure for the innovator's first project. Both of these equations were then analyzed using discriminant function analysis to determine whether they could identify variables which would differentiate function analysis to determine whether they could identify variables which would differentiate between innovators and non-innovators. The third part in the data analysis was the use of a separate discrim- inant function analysis to identify variables which differentiated between the known innovators (EDP project directors) and the known non-innovators. Finally, the data from faculty who had been involved in instructional development or instructional innovation projects 56 57 without EDP support were compared with the data from the innovators and the non-innovators to provide information on the effects of selection on the findings of this study. Three types of tables are used to present the data indicating how innovators differed from non-innovators. First, there are tables containing the results of univariate F-tests on the means of the measures used in this study; Tables 3, 6, and 9. These tables show which individual measures differed between innovators and non- innovators. Second, there are tables of the multiple correlations of the measures included in the regression equations or discriminant functions; Tables 1, 4, 7 and 10. The value of Multiple R or the square root of (l-Wilks lambda) in these tables shows how strong a relationship exists between the measures taken in combination and the value being predicted, e.g., number of EDP projects done or group membership; the value of Multiple R2 represents the cumulative percentage of the variance being explained when the measures listed to that point are included in the equation or function. The third type of table presents the coefficients for the measures to be included in either the regression equation or discriminant functions; Table 2, 5, 8, and 11. The results of this study will be presented in the following order. First, an analysis of the questionnaire returns will be reported. Second, the results from the attempt to use multiple regression equations to identify variables which differed for innovators and non-innovators will be described. Third, the results of the direct comparison of innovators and non-innovators using discriminant function analysis will be explained. Fourth, the results of comparisons of the unsupported innovators with both innovators and non-innovators will be 58 presented to indicate the effects of selection on the findings of this study. Finally, an analysis of the responses by the unsupported innovators and non-innovators to questionnaire items dealing with the effects of maturity will be reported. Questionnaire Returns Questionnaires were sent to 80 EDP project directors still on campus and to ten project directors who had left the campus because of retirement or new positions. Seventy-one of the on-campus group (89%) and eight of the off-campus group (80%) returned the questionnaires. From this group of 79 returns, 11 questionnaires had to be ommitted from the analysis since the respondents did not meet the criteria for inclusion in the study, i.e., were of instructor rank. This gave a return rate of 86% based on a maximum of 79 possible returns. The rank and academic disciplines of the faculty who did not respond were analyzed using Chi Square to determine whether there was systematic bias in the unreturned questionnaires along either of these dimensions. Academic discipline was used rather than academic department to ensure adequate cell sizes since there were so few respondents per department in the original sample. The test for academic ranks gave a Chi Square of 3.916 with two degrees of freedom. This was not statistically significant at p<.05. Therefore, there did not appear to be systematic bias with regard to rank or academic discipline. Questionnaires were also sent to 250 non-EDP supported faculty on campus. Returns were received from 180 faculty (72%). However, the total possible returns (250) was reduced to 232 to account for 11 faculty who were on sabbatical and one faculty member who was in the 59 hospital at the time of the study. Six of the returned questionnaires had to be ommitted since the respondents did not meet the criteria for inclusion in the study, i.e., did not teach at least two courses. Based on 232 possible returns, 75% were returned. The rank and departments of the faculty who did not respond were similarly analyzed using Chi Square to determine whether there was systematic bias in the unreturned questionnaire from either of these sources. The test for academic ranks gave a Chi Square value of 12.27 with two degrees of freedom. This was statistically significant at p<.Ol. There were more non—returns from professors and fewer from assistant professors than expected if there had been no bias. The test for departments gave a Chi Square value of 46.00 with 69 degrees of freedom which was not statistically significant at p<.05. Therefore, there was bias in the returns from the non-EDP supported faculty on the basis of rank but not on the basis of depart- ments. The effect from this bias should be minimal, however, since there were many individuals of each rank who did respond: 44 professors, 46 associate professors, and 82 assistant professors. The following sections report the results of various tests run on the data provided by the returned questionnaires. Summary data from the questionnaires returned is reported in Appendix C for the EDP gproject directors, Appendix D for the non-innovators, and Appendix E for the unsupported innovators. Regression Equation to Predict the Number of Projects A regression equation was generated to predict the number of PIijects done by an innovator which might identify a combination of measures which differed for innovators and non-innovators. Twenty-two equz foul don 2*: C1 of 60 of the 27 measures met the criteria for inclusion in the regression equation, and together they explained 67% (Multiple R2) of the variance found among the innovator group in terms of the number of EDP projects done. An analysis of variance of this value of Multiple R2 gave an F = 2.3657, with 22 and 26 degrees of freedom, which was statistically significant at p<.02. This statistical significance indicates that the relationship among the measures specified by the equation was not due strictly to chance. The correlation matrix used in the generation of this regression equation appears in Appendix C. Table 1 shows the multivariate correlations of each of the measures included in this regression equation. The value of Multiple R2 is the cumulative amount of variance accounted for by the measures combined with one another. This table also shows the value of the change in Multiple R2 associated with each measure and indicates which of these changes are statistically significant and not due strictly to change. Only the five measures shown in Table 1 which accounted for statistically significant changes in Multiple R2 were included in the equation used to compute regression scores on which classification of innovators and non-innovators would be based, since the effect of the rather variables on the equation was due to chance (p<.10). These five rmeasures accounted for 46% of the variance among innovators in terms of the number of EDP projects done (Multiple R2); their coefficients appear in Table 2. The 46% of the variance explained by these five measures gave an (Iverestimate of the amount of population variance explained. The amount of variance in the population explained after the appropriate 61 correction for shrinkage was only 40%. This regression equation, however, did not provide information on which measures and their related variables differed for innovators and non-innovators because when the regression scores for each individ- ual were computed and analyzed using discriminant function analysis, only 15 of the 68 known innovators (22%) were classified as innovators. The rest were classified as non-innovators. This indicates that Egg. measures in the regression equation did not differentiate between innovators and non-innovators even though_they did predict the number of EDngrojects done by an innovator. 62 Table 1 Multiple Correlations of Variables in Regression Equation to Predict the Number of Projects Multiple Multiple Multiple R2 Variable R R2 Change Knowledge .4944 .2444 .2444* Cosmopolite Style .5543 .3073 .0629* Opinion Leadership Change .6019 .3623 .0550* Age .6502 .4228 .0605* Innovativeness .6808 .4635 .0407* Local Information .6996 .4895 .0260 Teaching Percent .7074 .5004 .0109 Teaching Models .7216 .5207 .0203 Tenure .7317 .5354 .0147 Teaching Stability .7476 .5590 .0235 Resources .7551 .5702 .0112 Local Style .7623 .5811 .0109 Teaching Fatalism .7700 .5929 .0118 Student Fatalism .7770 .6038 .0109 Academic Discipline .7820 .6115 .0078 Cosmopolite Information .7869 .6192 .0077 Faculty Reaction .7899 .6239 .0047 Teaching Specialization .7924 .6280 .0041 Courses Taught .8000 .6399 .0120 Teaching Isolation .8153 .6648 .0249 Rank .8160 .6658 .0011 Participation .8166 .6669 .0010 *Significant p<.10 N = 49 63 Table 2 Unstandardized Regression Coefficients to Predict the Number of Projects 4 Variable Coefficient Knowledge .4493 Cosmopolite Style -.1007 Opinion Leadership Change .6182 Age -.0444 Innovativeness -.3502 (Constant) -2.0423 Regression Equation to Predict a Combined Success Score A regression equation was also generated to predict a combined success score for the first EDP project done by an innovator. Twenty- four of the 27 measures met the criteria for inclusion in this equation, and they accounted for 512 of the variance found among the innovator group. However, an analysis of variance gave an F of .9600, with 25 and 23 degrees of freedom, which was not statistically significant at p<.lO. Since this lack of significance indicates that the relation- ships among the measures specified by the regression equation was due primarily to chance, further analysis was not warranted. 64 Discriminant Function Analyeis to Discriminate Between Innovators and Non-innovators The discriminant function analysis provided both univariate and multivariate tests to discriminate between innovators and non- innovators. The univariate tests, F test between the means on each measure taken individually, will be presented first, then the multi- variate tests which indicated the relative importance of the variables taken in combination will be reported. The statistical significance of this discriminant function was established by a Chi Square test of Wilks lambda which gave a Chi Square value of 97.3869 with 24 degrees of freedom. This was statistically significant at p.<.OOl and indicates that the relation- ship among the variables specified by the function was not due to chance. The univariate F-tests identified 11 measures whose means differed for innovators and non-innovators. These measures represented the following individual variables: innovativeness, integration with social system, cosmopoliteness, information seeking about instruction, and opinion leadership; and the following perceptions of their departments: norms on innovation, norms on instructional strategies, and stability of instructional assignments. Multivariate analysis of the differences between these two groups indicated that all of these variables, with the exception of integration with the social system, were important in differentiating between innovators and non-innovators when the other variables were controlled for. It is worth noting that innovativeness and norms on instructional strategies accounted for 442 of the total variance explained. 65 This analysis also showed that specialization of teaching responsibility had an influence on the relationship among the other variables, even though it does not differ between innovators and non-innovators. Analysis of the directions of the differences between innovators, and non-innovators showed that EDP project directors, the innovators, had more innovativeness, more cosmopoliteness, more integration with the social systemi more information seeking about instruction, and more opinion leadership change than the non-innovators. The non- innovators, on the other handy_came from social systems which tended to more supportive of innovation and initially had more opinion leadership than the innovators. (Even though the results for integra- tion with the social system were mixed, use of local Information sources is a more direct operationalization than Years at MSU and provides a more statistically significant difference.) Table 3 shows the means of the measures for innovators and for non-innovators along with the values of the F-tests of the differences between the two means. The differences between the means of seven of the 27 measures were statistically significant at p<.01 and an additional 4 were statistically significant at p<.05. 66 Table 3 Univariate F Tests of Variable Means for Innovators and Non-innovators Variable Innovators Non-innovators Mean Mean F Teaching Models 4.37 6.41 43.4247** Innovativeness 1.60 2.74 37.4729** Cosmopolite Style 10.02 7.29 14.1159** Local Information 3.77 2.91 10.7379** Information Seeking 8.90 7.03 8.6778** Opinion Leadership Change 3.15 2.76 8.6635** Faculty Reaction 3.90 4.52 7.0071** Opinion Leadership .50 .71 5.0668* Years at MSU 9.71 13.00 5.0060* Cosmopolite Information 5.13 4.12 4.4886* Teaching Stability 3.29 3.95 3.9848* Resources 2.67 3.19 3.7565 Tenure .65 .81 3.5065 Age 41.40 44.98 3.3383 Local Style 7.38 8.40 3.3025 Knowledge 4.17 3.71 2.7101 Teacher Fatalism 2.88 3.34 2.4459 Courses Taught 3.65 4.19 2.2449 Teaching Isolation .27 .36 1.0802 Academic Discipline 1.94 1.79 .8635 Student Fatalism 3.54 3.81 .5855 Rank 4.10 4.22 .5733 67 Table 3 (continued) Variable Innovators Non-innovators Mean F Teaching Percent 59.17 56.55 .4130 Participation 5.73 5.86 .3518 Teaching Value 3.88 3.97 .0684 Teaching Importance 5.10 5.16 .0482 Teaching Specialization 7.69 7.66 .0031 *Significant p<.05 **Significant p<.01 N = 110 Degrees of Freedom 1, 108 The multivariate tests required that a discriminant function be generated to discriminate between innovators and non-innovators. Table 4 shows the multiple correlations of each measure with group membership coded as a dummy variable (innovator or non-innovator) resulting from this discriminant function. It also shows the statist- ical significance of the change in Multiple R2. Twenty-four of the 27 measures met the criteria for inclusion in this function, and together they explained 63% of the variance among innovators and non-innovators. The first eight measures in Table 4 provided incremental changes in Wilks lambda, which is related to Multiple R2 (Multiple R2=1 Wilks lambda), that were statistically significant at p<.Ol indicating that their contribution to the function was not due to chance. In combination, these eight measures accounted for 60% of the variance between innovators and non-innovators. It was inappropriate to 68 (:omsider the three additional measures which provided incremental cihanges in Wilks lambda significant at only p<05 since teaching especialization was highly correlated with teaching isolation (.86). frhis leads to problems of intercorrelation in the analysis. Further- ‘more, these three measures only accounted for an additional 2% of the variance. The amount of variance in the sample explained by the eight measures which provided statistically significant (p<.01) incremental changes in Wilks lambda gave an overestimate of the population variance. The amount of variance in the population explained after correction for shrinkage was 572. Table 4 Multiple Correlations Among Variables in Discriminant Function to Discriminate Between Innovators and Non-innovators Variable Wilks Multiple R2 Multiple R2 lambda (1-1ambda) Change Teaching Models .7132 .2868 .2868** Innovativeness .5599 .4401 .1533** Information Seeking .5207 .4793 .0392** Opinion Leadership .4824 .5176 .0383**4 Opinion Leadership Change .4471 .5529 .0353** Teaching Stability .4336 .5664 .0135** Teaching Isolation .4164 .5836 .0172** Teaching Importance .4027 .5973 .0137** Teaching Specialization .3941 .6059 .0086* Cosmopolite Style .3858 .6142 .0083* 69 Table 4 (continued) Wilks Multiple R2 Multiple R2 Variable lambda (l-lambda) Change Faculty Reaction .3804 .6196 .0054* Local Style .3767 .6233 .0037 Courses Taught .3723 .6277 .0043 Academic Discipline .3710 .6290 .0013 Teaching Value .3700 .6301 .0010 Teaching Percent .3692 .6308 .0008 Cosmopolite Information .3685 .6315 .0007 Student Fatalism .3679 .6321 .0006 Teacher Fatalism .3673 .6327 .0006 Resources .3669 .6331 .0004 Participation .3667 .6333 .0002 Knowledge .3666 .6334 .0001 Age .3665 .6335 .0001 .Rank .3664 .6336 .0001 *Significant p<.05 **Significant p<.01 N = 110 Table 5 shows the discriminant function coefficients for these eight measures when they were the only measures in the function. Since these coefficients are standardized, they represent the relative contribution of each variable to the function, and can be used to compare the contributions among the measures, controlling for the other variables. 70 To test the ability of these analyses to differentiate between innovators and non-innovators, the classification coefficients shown in Table 5 were used in two functions to compute classification scores for each individual. The classification function which yielded the highest score determined the group to which an individual would be assigned. These classification functions classified 56 of the 68 innovators (82%) and 10 of the 88 non-innovators (11%) as innovators. The remaining faculty were classified as non-innovators. Discriminant Function Analysis of the Innovators Not Supported by EDP To determine the impact of selection on the differences found between innovators and non-innovators, five tests were made on the data from those faculty who reported being involved in instructional development or instructional innovation projects without EDP support. These tests involved: using the differences found between innovators and non-innovators represented by the classification coefficients from Table 5 to classify these unsupported innovators as either innovators or non-innovators, using discriminant function analysis to provide univariate and multivariate tests comparing these unsupported innovators with those innovators supported by EDP, and univariate and multivariate tests comparing these unsupported innovators with the non-innovators. Classifying_phe unsupported innovators on the basis of the differences between innovators and non-innovators. Using the differences found between innovators and non-innovators (as represented by the classification coefficients in Table 5) would provide information on the effects of selection on the findings. If the unsupported 71 innovators were classified as innovators, the effects of selection would be small. If the unsupported innovators were not classified as innovators, the effects of selection would be important in the interpretation of the findings from this study. When the classification coefficients from Table 5 were used to classify the unsupported innovators as either innovators or non-innovators, only 22 of the 86 (26%) were classified as innovators. This compares with the 82% of the EDP project directors which were classified as innovators. Therefore, the unsupported innovators are not like the EDP supported innovators in terms of the eight measures in Table 5. Table 5 Discriminant and Classification Coefficients to Discriminate between Innovators and Non-innovators Standardized Unstandardized Discriminant Classification Function Function Variable Coefficient Coefficient Innovators Non-innovators Teaching Importance .2504 2.9258 3.3576 Teaching Models .8829 1.4208 2.5333 Innovativeness .5599 3.7053 4.9012 Teaching Stability .3891 1.7878 2.3238 Teaching Isolation .3346 5.1766 6.9050 Opinion Leadership Change -.4046 7.0328 5.6883 Information Seeking -.5445 .5846 .2018 Opinion Leadership .4000 -2.8198 - .8473 (Constant) -30.l38 -37.590 72 Comparison with EDP-suppprted innovators. The differences between the EDP-supported innovators and the unsupported innovators could be identified using discriminant function analysis in the same manner was done to identify the differences between innovators and non-innovators. This analysis provided both univariate F-tests between means and multivariate tests. The multivariate tests indicated which measures were most important in differentiating between EDP- supported innovators and unsupported innovators, when the other measures were controlled for. The univariate F-tests identified seven measures whose means differed between the two groups. These measures represented the following variables: fatalism, innovativeness, cosompoliteness, opinion leadership, norms on instructional strategies, resources for instructional improvement, and stability of instructional assignments. Multivariate analysis of the differences between these two groups indicated that fatalism, innovativeness, opinion leadership and the perception of norms on instructional strategies were important in differentiating between EDP-supported innovators and unsupported innovators, when other variables were controlled for. Differences in the norms on instructional strategies were most important in differentiating between the two groups of innovators (accounting for 28% of the total variance), as was the case in differentiating between innovators and non-innovators (where it accounted for 29% of the total variance). This analysis also showed that two individual variables, information seeking about instruction and knowledge about instructional innovations, had an influence on the relationship among the other variables, even though they did not 73 differ between the two types of innovators. Analysis of the directions of the differences between EDP- supported innovators and non-innovators showed that EDP-supported innovators had more innovativeness and more cosmopoliteness than the unsupported innovators. The unsupported innovators were more fatalistic, had more opinion leadership; and came from departments which tended to be more supportive of innovation. The results of the univariate F-tests are shown in Table 6. Seven measures were identified with statistically significant differences between the means at p<.05. Of these seven, four were significant at p<.01. Table 6 Significant Univariate F Tests of Variable Means for EDP-Innovators and Unsupported Innovators Unsupported EDP-Innovators Innovators Variable Mean Mean F Teaching Models 4.37 6.56 43.9152** Opinion Leadership .50 .83 15.4371** Innovativeness 1.60 2.13 10.1344** Resources 2.67 3.38 7.2560** Cosmopolite Style 10.02 8.25 6.4511* Teacher Fatalism 2.88 3.52 4.8527* Teaching Stability 3.29 3.98 4.5331* *Significant at p<.05 **Significant at p<.01 N = 115 Degrees of Freedom 1, 113 74 The multivariate tests comparing EDP-supported innovators with unsupported innovators were conducted using discriminant function analysis. A discriminant function was generated which explained 51% of the variance between these two groups of innovators and gave a Chi Square that was statistically significant at p<.OOl, indicating the relationship among the variables specified by the function was not due to chance. Only the first seven measures included in this function, as shown in Table 7, provided an incremental change in Wilks lambda significant at p<.05. These seven measures accounted for 44% of the total variance, and were the only measures used in the classification functions derived from the discriminant scores, since the contributions of the other variables were not significantly different from chance. The discriminant and classification coefficients for these variables are shown in Table 8. The classification functions classified 46 of the 68 EDP- supported innovators (68%) and 14 of the 70 unsupported innovators (20%) as innovators. When the non-innovators were classified using these functions, only 5 of the 90 non-innovators (6%) were classified as innovators. All others were classified as non-innovators. Even though these classification functions were less able to correctly classify innovators and non-innovators than those based on Table 5 using the discriminant function analysis of EDP-innovators and non-innovators, they do indicate that the differences found between EDP-supported innovators and non-innovators were valid, since a large proportion of each group could be differentiated from the other group. 75 Table 7 Significant Multiple Correlations Among Variables in the Discriminant Function to Discriminate Between EDP-Innovators and Unsupported Innovators Wilks Multiple R2 ‘Multiple R2 Variable lambda (l-lambda) Change Teaching Models .7201 .2799 .2799** Opinion Leadership .6656 .3344 .0545** Innovativeness .6337 .3663 0319** Opinion Leadership Change .6115 .3885 .0222** Information Seeking .5852 .4148 .0263** Student Fatalism .5714 .4286 .0138* Knowledge .5571 .4429 .0143* *Significant at p<.05 **Significant at p<.Ol N - 115 76 Table 8 Discriminant and Classification Coefficients to Discriminate Between EDP-Supported Innovators and Unsupported Innovators Standardized Unstandardized Discriminant Classification Function Function Variables Coefficients Coefficients EDP-Supported Non-Supported Innovators Innovators Knowledge -.2537 1.7170 1.4024 Teaching Models .9840 .7655 1.6098 Innovativeness .2584 2.7102 3.2064 Student Fatalism .2757 1.3336 1.6001 Opinion Leadership Change -.3601 7.2255 6.3154 Information Seeking -.2443 .8378 .7147 Opinion Leadership .5827 -l.7l4l .4915 (Constant) -24.471 -27.292 77 Comparison with non-innovators. Just as was done in the previous analyses, the differences between the unsupported innovators and the non-innovators could be identified using discriminant function analysis. In this case, only four measures whose means differed between the two groups were identified by the univariate F-tests (see Table 9). The measures represented the following variables: innovativeness, information seeking about instruction, integration with the social system, and opinion leadership. Table 9 Significant Univariate F Tests of Variable Means for Unsupported Innovators and Non-innovators Unsupported Innovators Non-innovators Variable Mean Mean F Innovativeness 2.13 2.74 10.93044** Information Seeking 8.94 7.03 10.2393** Local Information 3.75 2.91 10.6138** Opinion Leadership Change 3.02 2.76 3.9899* *Significant at p<.05 **Significant at p<.Ol N - 121 Degrees of Freedom 1, 119 The multivariate analysis between the two groups provided a discriminant function that gave a Chi Square which was statistically significant at p<.001 and indicated that innovatives, integration with the social system, and opinion leadership were important in differentiating 78 between the unsupported innovators and the non-innovators when the other variables were controlled for. However, it should be pointed out that only 30% of the variance between the two groups was explained using all the measures in this study; and, when only those four variables which provided a statistically significant change in Multiple R2 (p<.05) were considered, only 19% of the total variance was explained. The results of this multivariate analysis are shown in Table 10. I Table 10 also shows that in addition to innovativeness, integra- tion with the social system, and opinion leadership, specialization of teaching responsibility had an influence on differentiating between the two groups even though it did not differ between the groups. Table 10 Significant Multiple Correlations Among variables in the Discriminant Function to Discriminate Between Unsupported Innovators and Non-innovators Wilks Multiple R2 Multiple 112 Variable lambda (l-lambda) Change Innovativeness .9159 .0841 .0841** Local Information .8595 .1405 .0564** Teaching Isolation .8360 .1640 .0235* Opinion Leadership Change .8145 .1855 .0215* *Significant at p<.05 **Significant at p<.Ol N - 121 79 Analysis of the directions of the differences between the unsupported innovators and the non-innovators showed that the supported innovators had more innovativenessy_more information seekingyubout instruction, more integration with the social systeug and more opinion leadership than the non-innovators. When the four measures providing statistically significant changes in Multiple R2 were used in the classification functions, 60 of the 84 unsupported innovators (71%) and 31 of the 88 non-innovators (35%) were classified as innovators. When the EDP-supported innovators were classified using these functions, 51 of the 68 EDP-innovators (75%) were classified as innovators. All others were classified as non- innovators. The discriminant and classification coefficients for these measures are shown in Table 11. Since over 70% of both the EDP-supported innovators and the unsupported innovators were classified as innovators, and only 35% of the non-innovators were classified as innovators, it shows that the two grpups of innovators are more like each other than they are like non—innovators. 80 Table 11 Discriminant and Classification Coefficients to Discriminate Between Unsupported Innovators and Non-innovators Standardized Unstandardized Discriminant Classification Function Function Variables Coefficients Coefficients Unsupported Innovators Non-innovators Innovativeness .5982 2.6659 3.1991 Teaching Isolation .4290 - .3887 .4897 Opinion Leadership Change -.3894 6.2498 5.7468 Local Information -.6391 2.0453 1.6303 (Constant) -16.04l -l4.776 Effects of Maturation The effects of maturation on the findings of differences between innovators and non-innovators were indicated by analyzing the responses to four items provided by the non-EDP supported faculty. These items asked for a comparison of the individual's current levels of activity or current conditions with the levels five years previously. Table 12 shows the responses to these items. In all categories, the predominant response was that activities or conditions were the same as five years previously. A Chi Square test of the percentage of respondents indicating a difference (in either direction) was not statistically significant at p<.05. Therefore, the same overall amount of change was reported for all four categories. This test is shown in Table 13. 81 Table 12 Comparison of Current Activities with Five Years Previously Activity/Condition Amount Compared with Five Years Ago its ear-e; 112.22 Information Seeking 27(162) 96(57%) 46(272) Teaching Importance 24(l4%) 102(6l%) 41(25%) Participation in Department l3(8%) 126(75%) 28(17%) Positive Reaction of Faculty to Innovation 23(l4%) 110(662) 33(20%) Table 13 Chi Square Test of Percent of Faculty Reporting Change in Current Activities Activity/Condition Expected % Observed % Information Seeking 35.25 43 Teaching Importance 35.25 39 Participation in Department 35.25 25 Positive Reaction to Faculty 35.25 34 to Innovation Chi Square - 5.13 When the total changes (in either direction for all categories was computed for each respondent, the mean number of activities or conditions which had changed for each respondent was 1.43 with a standard deviation of 1.00. 82 These findings suggest that while there has been some change during the past five years, it has not been in any one direction (more or less of an activity or condition), nor has it appeared to be sufficient to base the findings of this study on the effects of maturation. Majpr Findiugg The data analysis showed that eight of the 18 characteristics studied differed between innovators and non-innovators. It was found that: l. Innovators differed from non-innovators by showing: a. Greater innovativeness b. More integration with the social system c. Greater cosmopoliteness d. Greater information seeking about instruction e. Greater Opinion leadership change 2. Innovators' perceptions of their departments differed from the perceptions of non-innovators by showing: a. Less supportive norms on innovativeness b. Less supportive norms on instructional strategies c. More stability of instructional assignments An additional major finding was that there were three distinct groups of faculty: innovators, unsupported innovators, and non- innovators. Summary Identifying variables which differed for innovators and non-innovators using multiple regression equations to predict the number of EDP projects done by an innovator or a combined score of 83 project success did not prove useful. While a statistically significant regression equation was generated to predict the number of EDP projects done, the relationships among the variables represented in it could not be used to differentiate between innovators and non-innovators. It was not possible to generate a statistically significant regression equation to predict a combined score of project success. Discriminant function analysis, howevep, identified eigup characteristics which differed for innovators and non-innovators. These characteristics were: innovativeness, integration with the social system, cosmopoliteness, information seeking about instruction, opinion leadership, norms on innovation, norms on instructional strategies, and stability of instructional assignments. The innovators had more innovativeness, cosmopoliteness, integration with the social system, information seeking about instruction, and opinion leadership change than the non-innovators. When the unsupported innovators were compared with both the EDP-supported innovators and the non-innovators, it was found that there were fewer differences between the unsupported innovators and the non-innovators than between the unsupported innovators and the EDP-supported innovators. The effect of maturation due to the difference in time periods on which responses were based appeared to be small since most of the non-EDP supported faculty indicated that the current levels of activity or conditions were about the same as five years previously. 84 The specific measures identified by discriminant function analysis whose means were significantly different for innovators and non-innovators (p<.05) were: Teaching Models, Innovativeness, Cosmopolite Style, Local Information, Information Seeking, Opinion Leadership Change, Faculty Reaction, Opinion Leadership, Years at MSU, Cosmopolite Information, and Teaching Stability. Eight measures were also identified which, in combination, explained 60% of the variance between innovators and non-innovators in the sample. This represents 57% of the variance in the population. These eight measures were: Teaching Models, Innovativeness, Information Seeking, Opinion Leadership, Opinion Leadership Change, Teaching Stability, Teaching Isolation, and Teaching Importance. The classification functions using these eight measures correctly classified 82% of the innovators and 89% of the non-innovators. EDP-supported innovators were also compared with unsupported innovators to identify the effects of selection on the findings. Seven measures were identified whose means were significantly different between the two groups of innovators (p<.05). These seven measures were: Teaching Models, Opinion Leadership, Innovativeness, Resources, Cosmopolite Style, Teacher Fatalism, and Teaching Stability. In addition, seven measures were identified which, in combination explained 44% of the variance between EDP-supported and unsupported innovators. These measures were: Teaching Models, Opinion Leadership, Innovativeness, Opinion Leadership Change, Information Seeking, Student Fatalism, and Knowledge. There were fewer measures that differed between unsupported innovators and non-innovators than between unsupported innovators and 85 EDP-supported innovators. Only four had means that were significantly different between unsupported innovators and non-innovators (p<.05): Innovativeness, Information Seeking, Local Information, and Opinion Leadership Change. Only four measures produced significant incremental changes in Wilks lambda (p<.05); in combination they explained 19% of the variance. These four measures were Innovativeness, Local Informa- tion, Teaching Isolation, and Opinion Leadership Change. CHAPTER V SUMMARY AND DISCUSSION Introduction The purpose of this study was two-fold: first, it was to identify how innovators differed from.non-innovators, and second, it was to identify how innovators' perceptions of their departments differed from those of non-innovators. These differences were to be examined in terms of 27 measures representing 18 characteristics selected from the Rogers and Shoemaker (1971) model on the diffusion and adoption of innovations. Questionnaires were sent to a sample of 90 EDP project directors, the innovators, and 250 non-EDP supported faculty, the unsupported innovators and the non-innovators. Data from the questionnaires were analyzed using multiple regression and discriminant function analysis to identify those variables which differed, individually or in combination with other variables, between innovators and non-innovators. Of the two types of analysis used in this study, only discriminant function analysis was able to identify characteristics which differed between innovators and non-innovators. The multiple regression equation to predict the number of projects done by an individual was statistically significant, but was not useful for identifying differences between innovators and non-innovators. It was not possible to generate a statistically significant multiple regression equation to predict a combined measure of project success 86 87 for an individual's first EDP project. The discriminant function analyses identified eight of the 18 characteristics from the Rogers and Shoemaker model which differentiated between innovators and non-innovators, providing answers to the two research questions raised in Chapter III. Furthermore, it was discovered that the two types of innovators (EDP-supported and unsupported) were not alike. The differences between the two groups of innovators were consistent with Rogers' and Shoemaker's distinctions between innovators and early adopters. In this study, the EDP-supported innovators had the characteristics of innovators, while the unsupported innovators had the characteristics of the early adopters. As a result of the analyses done in this study, four conclusions were reached: 1. Eight differences can be identified between innovators and non-innovators in terms of personal and social system characteristics. Three groups of faculty can be identified: innovators, early adopters, and non-innovators. Innovators perceive that their departments §2_22£ provide sufficient financial and/or psychological support for instructional innovation. Only a portion of the individual and social system characteristics from the Rogers and Shoemaker model on the diffusion and adaption of innovations appears generalizable to higher education settings. Each of these four conclusions will be dealt with separately in this chapter. The data and rationale supporting each conclusion will be 88 presented, along with a discussion of its implications. Conclusion No. l Eight differences can be identified between innovators and non-innovators in terms of personal and social system characteristics. Discriminant function analysis of the data from this study identified eight of the 18 characteristics from the Rogers and Shoemaker model which differentiated between innovators and non-innovators. This provided answers to the research questions raised in Chapter III. The data analysis indicated that: l. Innovators differed from non-innovators by showing: a. Greater innovativeness b. Greater integration with the social system c. Greater cosmopoliteness d. Greater information seeking about instruction e. Greater opinion leadership change 2. Innovators' perceptions of their departments differed from the perceptions of non-innovators by showing: a. Less supportive norms on innovativeness b. Less supportive norms on instructional strategies c. More stability of instructional assignments The multivariate analysis of the data indicated that the two most important characteristics in differentiating between innovators and non-innovators were innovativeness and norms on instructional strategies. Together they accounted for 44% of the total variance between innovators and non-innovators when the other variables were controlled for. 89 An indication of the usefulness of the differences identified between innovators and non-innovators was also provided by the multi- variate analysis. This analysis made it possible to combine the eight characteristics which differed between innovators and non- innovators to classify the respondents into the two groups, as well as classify the group of unsupported innovators (early adopters). This classification provided a test of the validity of the findings. The test correctly classified 82% of the innovators, 89% of the non- innovators, while 74% of those who claimed to be innovators, without EDP support, were classified in a single group (as it turns out, they were classified as non-innovators). There may have been other differences between innovators and non-innovators which went unidentified because of two reasons: Imprecision of the measures used and homgeneity of the university. For example, the measures of teaching specialization were ambiguous and partially misleading since they did not ask the respondent to provide a detailed response. The measure of participation asked for a single self-report on the sum of all types of participation in the department and was too general to provide differentiation between respondents. The selection of academic discipline as the measure of membership in a modern and/or integrated system appeared to be inappropriate since it did not differ for innovators and non-innovators while other social system variables did, i.e., the variables dealing with social system norms. The measures of fatalism may also have been to imprecise to identify differences among the groups. While fatalism differed between innovators and early adopters, different fatalism measures were signi- ficant in the univariate and multivariate tests. Since only two 90 fatalistic statements were used, the various dimensions of fatalism may not have adequately been measured. It is also possible, however, that the non-innovators did not differ in fatalism from the innovators, but for other reasons (e.g., support and resources from the social system) chose not to innovate. Homogeneous characteristics of the university may have combined with the measures of several variables to mask differences between innovators and non-innovators. For example, the university may be fairly homogeneous with regard to size of teaching load and norms on the importance of teaching. No differences were found in this study; though, had more sensitive measures been used which indicated differences between innovators and non-innovators, these differences may not really have been large enough to be meaningful. In a similar fashion, the values of the faculty may be homogeneous with regard to knowledge about instructional innovations and the importance of status. Respondents may not have been able to clearly identify how much they actually knew about instructional innovations since they may have perceived that they were expected to know a great deal while the measure used in this study did not provide a standard with which a comparison could be made. The lack of difference in terms of status may reflect another characteristic of the university: the university may not place great value on status as measured by tenure or academic rank. There were mixed results with respect to integration with the social system. Non-innovators had more years at MSU; yet, innovators used more Local Information sources. The exact relationship of Years at MSU with integration with the social system is not clear. 91 It was intended to represent the potential for becoming integrated, though it potentially could have also represented the potential for forming cosmopolite sources of information, being affected by the norms of the social system, becoming an opinion leader, etc. Use of local information sources, however, is a more direct measure of integration with the social system; and since it represented a more statistically significant difference between innovators and non- innovators, it was used as the measure of integration with the social system. Even though differences were found for only eight characteristics, this study was able to account for 60% of the variance within the sample of innovators and non-innovators (which is equivalent to 57% of the variance in the population). Rogers and Shoemaker (1971, pp. 191-192) summarize 36 multiple correlation studies of innovativeness, primarily involving farmers in the U.S. and abroad (none dealt with innovation in educational systems). Of these 36 studies, only eight were able to explain more variance. Four of these studies were able to explain this amount of variance with eight or fewer variables. One study explained 81% of the variance using four variables; however, the other three studies only explained approximately 64% of the variance using five or six variables. The three studies that used more variables explained from 68.9% to 87.5% of the variance using from 27 to 51 variables. Implications of conclusion 1. There are two implications of being able to identify these eight differences between innovators and non-innovators. First, this information can be used to predict whether an individual is more likely to be an innovator, early adapter, or a 92 non-innovator. Second, this information can be used to devise strategies to increase the adoption of innovations. Using the procedures described in Chapter III for combining the eight characteristics into classification functions, individuals can be identified as more likely to be innovators, early adopters, or non-innovators. The appropriate values for the variables in these classification functions can be obtained by having the individual respond to the measures appropriate to teach variable as indicated on the questionnaires in Appendices A and B. Identifying whether an individual is more likely an innovator, early adopter, or a non-innovator should be useful in program evaluation and management because it makes it possible to determine whether effort and resources are being spent on the target audience- innovators, early adopters, or non-innovators. Since early adopters are more integrated with their social system than innovators, it could also be useful to differentiate between these two groups. It is also possible to use the information about how innovators differ from non-innovators to devise strategies to increase the adoption of innovations. This use is based on two assumptions which need further testing. The first assumption is that there is a cause- effect relationship between the variables identified in this study as differentiating between innovators and non-innovators and an individual becoming an innovator or early adopter. The second assumption is that making individuals more like innovators or early adopters will increase the adoption of innovations. The strategies to increase the adaption of innovations attempt to make everyone's characteristics similar to the characteristics of 93 innovators or early adopters in terms of the variables in this study which differed between innovators or early adopters and non-innovators. For example, since innovators sought more information about teaching and learning than non-innovators, a strategy might be devised which rewards or encourages information seeking by individual faculty. Another strategy might be to make information seeking easier by increasing the number of magazines or newsletters circulated to faculty. Conclusion No. 2 Three groups of faculty can be identified: innovators, early adopters, and non-innovators. Discriminant function analysis comparing the unsupported innovators with both the EDP-supported innovators and the non-innovators indicated that the two groups of innovators were not alike. These two groups differed in terms of fatalism, innovativeness, cosmopoliteness, opinion leadership. With respect to perceptions of their departments, the innovators differed from the early adopters by showing less supportive norms on instructional strategies, fewer resources for instructional improvement, and more stability of instructional assignments. These differences were confirmed by the results of the classification of respondents by means of the multivariate analysis. This classification assigned 68% of the EDP-supported innovators to one group and 80% of the unsupported innovators to a second group. The differences between the two groups of innovators can be explained in terms of differences Rogers and Shoemaker identify between innovators and early adopters. The description of innovators indicates that they are more innovative and cosmopolite, seek more 94 information, and are less a part of their social system than other adOpters (1971, p. 183). Early adopters are described by Rogers and Shoemaker as: . . . more integrated a part of the local social system than are innovators. . . . This adopter category, more than any other, has the greatest degree of opinion leader- ship in most social systems. . . . (p. 184) The unsupported innovator appears more integrated in his social system, at least in terms of opinion leadership which is related to such integration according to Rogers and Shoemaker, and that social system appears to provide support for that individual in terms of both norms and resources which support innovation. It may be the interaction of these factors (which differ between the EDP-supported innovators and the unsupported innovators) with the two other important factors (information seeking about instruction and knowledge about instructional innovations) which keep the "early adopters" (unsupported innovators) from being innovators. Since the early adopters were not identical with the innovators, the early adopters were compared to the non-innovators using discriminant function analysis. This analysis indicated that only four of the variables used in this study differed between early adopters and non- innovators: innovativeness, information seeking about instruction, integration with the social system, and opinion leadership. The early adopters had more of each of these characteristics. This analysis also indicated that information seeking about instruction was not an important difference, but that an additional variable, specialization of teaching responsibility, was. This latter variable represents in part, the relationship between the individual and the social system (an individual who is not specialized may share 95 teaching responsibilities for certain courses with others in the social system). The differences found between the early adopters and the non- innovators further support identification of the unsupported innovators as early adopters rather than innovators. The early adopters differed from the non-innovators on four of the same variables that differed between the innovators and the non-innovators-—the most important of these, as indicated by the multivariate analysis, being innovativeness. Furthermore, the additional variable which entered into this multivariate relationship between non-innovators and early adopters involved the social system, while the additional variables in the multivariate relationship between innovators and early adopters involved personal characteristics. This is consistent with the Rogers and Shoemaker generalization that social system effects may be more important in,” explaining the differences between adopter categories than individual characteristics of the adopters (see the discussion of Rogers' and Shoemaker's generalizations in Chapter II). In this case, there is a greater difference between the non-innovator category and the innovators or early adOpters than between the two caregories of innovators (innovators and early adopters.) Multivariate analysis showed that over 70% of both the innovators and the early adopters could be classified in the same group and 65% of the non-innovators classified in a separate group. Combined with the classification results, this confirms the existence of three distinct groups of faculty: innovators, early adopters and non—innovators. Implications of conclusion 2. One of the implications of 96 Conclusion 2 is that it provides a cross-check for identifying whether an individual is an innovator and less integrated in his social system, an early adopter and more highly integrated in his social system (in terms of opinion leadership), or a non-innovator. There is a second implication of this conclusion: an educational development program based on Optional decisions to adapt innovations does not reach a wide cross section of the faculty. In Chapter I it was suggested that all faculty might not choose to participate in an educational development program which was based on optional decisions to adopt innovations. (This is the type of EDP activity in effect at MSU.) It turns out that a program of this type, which relies on the individual faculty member to seek it out for support, did not reach a wide cross section of the faculty since it generally did not reach those faculty who were most integrated in their social systems and/or who were opinion leaders, and it did not reach those least integrated in their social system——the non—innovators. Therefore, the effects of the projects in such a program would be less likely to influence faculty members not involved in the project and would be less likely to spread rapidly throughout the institution-- especially to those not well integrated within that institution. It should be pointed out, however, that there are indications that the opinion leadership of an innovator increases following the adoption of an innovation. This change suggests that over a longer period of time the innovators would have an influence on their colleagues. The nature of this influence and how long it takes to develOp remain unclear. A third implication of this conclusion focuses on ways to 97 increase the adoption of innovations. Two of the ways to increase this adoption are: 1) make the individuals more like innovators, and 2) make the individuals more like early adopters (since both innovators and early adopters adopt innovations more readily than non-innovators). The implication from this conclusion is that it is not always necessary to make an individual like an innovator. When it is possible to change the social system's norms and resources to more supportive of innovation, adaption of innovations will probably be more widespread and will not require outside support. In this case, EDP effort should be directed to making individuals more like early adopters--increasing their opinion leadership, integration with the social system, cosmopoliteness, etc.--than like innovators since innovators are less integrated in their social system and would be less influenced by changes within it. Where it is not possible to make the necessary changes in the social system or where it is more important to have a few individuals adopt innovations quickly, maintaining the individual's integration with the social system would not be as important. Therefore, making the individual more like an innovator is appropriate. This can lead to faster adoption of innovations, but that adoption would not be as widespread. Such a strategy might be appropriate in situations where one step in changing the norms of the social system in terms of instructional strategies required that new instructional strategies be implemented within the social system. In such a case, making several faculty innovators might be a quick way to introduce these new instructional strategies. 98 Conclusion No. 3 Innovators perceive that their departments do not provide sufficient financial and/or psychological support for instructional innovation. This implies that both financial and psychological support are necessary for the adoption of innovations to occur. The innovators had little or no psychological support from within their social systems since the norms of the social systems did not support instruc- tional innovation, and they were apparently not recognized as opinion leaders about instruction. Need for this type of support, however, may account for the innovators being more cosmopolite--making use of outside reference groups for support. Financial support was, of course, provided by EDP. On the other hand, the early adopters came from social systems whose norms and resources supported instructional innovation. Further evidence of the importance of financial and psychological support comes from the multivariate analysis which showed that these innovators were more integrated with their social systems than non- innovators, but suggests that being integrated in a social system.which ‘may_have provided psychological support but did not provide resources was not sufficient to bring about the actual adoption of innovations without outside funds. The most important variables in differentiating between innovators and non-innovators were innovativeness and norms on instructional strategies. A difference in terms of innovativeness merely provides an additional test of validity for this study since it is a measure of when an individual generally adapts innovations 99 compared to his colleagues. By this definition, innovators in study would have to be more innovative than non-innovators if the groups were correctly selected to represent these two conditions. The importance of norms on instructional strategies is that it represents the influence of the social system on the individual. To innovate in a system which did not have or support a variety of instructional strategies, the innovator had to seek outside financial support (from EDP), be more cosmopolite, and seek more information about instruction. Implications of conclusion 3. The implication of this conclusion is that to increase the adOption of innovations it is necessary to insure that individuals perceive that there is both financial and psychological support. If this support cannot be obtained from the individual's social systems, it must be provided from outside--using outside funding sources and increasing the cosmopoliteness of these individuals. Conclusion No. 4 Only a portion of the individual and social system character- istics from the Rogers and Shoemaker model on the diffusion and adoption of innovations appearngeneralizable to higher education settings; The findings from this study can be compared to prior research (as discussed in Chapter II) from Rogers' and Shoemaker's generalizations from research in many fields and in terms of prior research from higher education alone. The findings from prior research and from this study will be compared in a side-by-side format to allow easier comparisons. 100 This comparison shows that there are a number of major discrepancies with the findings from the research in higher education; though this is primarily because there is so little of it. It also appears that only a portion of the rogers and Shoemaker model is appropriate for higher education. 'Rggers' and Shoemaker's Generali- This research. zations. Differences between 17 of the 18 Differences on eight of the 18 characteristics were found between characteristics were found earlier adopters and later between innovators and non- adopters (no difference was innovators; and on seven of the found for age). 18 characteristics between innovators and early adopters. A maximum of 882 of the variance This study explained 60% of was explained, though most the variance. studies explained less. Earlier adopters had more Innovators had less of all these opinion leadership, positive characteristics than non- norms on innovation and innovators, and innovators led instructional strategies. early adopters only in terms of cosmopoliteness and innovative- ness. (All other differences were in the direction indica- ted by the model.) This may be due to the innovators' perceiving their social systems as traditional and unsupportive of change. This study found direct support for those Rogers and Shoemaker generalizations dealing with innovativeness, cosmopoliteness, information seeking, and opinion leadership when innovators and early adopters were compared to non-innovators. There were mixed results with respect to the generalization on integration with the social system but they appeared to support the model. The results on social system characteristics were in the opposite direction from that predicted by the model. 101 An important reason why only a portion of the Rogers and Shoemaker model was supported by this study is that the cultures on which each was based were different. The Rogers and Shoemaker model was based primarily on data from agricultural innovation studies which were frequently done in less developed or primitive cultures. This study was done in a more modern culture, higher education, which has many differences with the agricultural setting, e.g., level of abstractions dealt with, types of rewards available, ability to delay gratification, etc. Therefore, part of the Rogers and Shoemaker model does not apply in a higher education setting. There is little overlap or agreement between these findings and the findings of past research from higher education (only the findings with respect to age are consistent among higher education studies). The lack of agreement among the various higher education studies is probably due to the different definitions of "adapters", use of different operational definitions and measures for the variables studied, and different types of innovations studied. Prior research from hi er education. This research. Innovators and non- innovatars were no different in age, but non-innovators had more years at MSU. Adapters and non-adopters were no different in age and had no difference in years at the institution. Adapters were less fatalistic Innovators and non-innovators than nan-adopters. did not differ on fatalism, but innovators were less fatalistic than early adopters. Adapters were more integrated The results were mixed. with their social system than Innovators had fewer years at nan-adopters. MSU, but made more use of local information. This may be due to different operational definitions of integration with the social system. 102 adapters had smaller teaching No difference in teaching loads; however, these results loads between innovators were questionable. and non-innovators. Differences existed between No differences between academic disciplines. academic disciplines. The differences observed by prior research were on characteristics not included in this study. Implications of conclusion 4. There are two implications which result from this conclusion. First, a comprehensive theory is needed to explain how the variables in the model affect the diffusion and adoption of innovations. Second, the results from past research cannot be used to predict which portions of the model which have not been tested are most likely to apply to higher education. There is not a comprehensive theory which explains why the variables in the Rogers and Shoemaker model fit together. Currently, the primary study and discussion has been on the relationship between the various variables and the adaption of innovations. Little has been presented on the interaction of the variables in the model and how these variables act to "cause" or "inhibit" the adoption of innovations. This makes it very difficult to interpret why this study did not conform to the entire model in this particular setting. Furthermore, since only 452 of the variables from the model which were tested in this study showed a difference between the innovators and the nan-innovators, and since there is not a comprehen- sive theory to use in explaining why this study did not conform to the entire model, it is difficult to predict which of the other variables from the Rogers and Shoemaker model would apply to higher education. Therefore, it is difficult to draw on the Rogers and Shoemaker generalizations dealing with these other variables in the 103 design of strategies to increase the adoption of innovations. Future Research Three areas of future research are indicated by the results of this study. First, other variables from the Rogers and Shoemaker model should be examined. Second, the effect of EDP support on a department over time should be studied. Third, this study should be replicated with better controls. Research including other variables. This study omitted communication and consequences variables from the Rogers and Shoemaker model. It also omitted several of the individual and social system variables--especially those relating to attitudes. Further examination of the model is warranted since this study found less than 45% (eight of the 18 characteristics tested) of the model to be applicable to higher education. Effect of EDP support over time. EDP support may have more effect on the social system than indicated by this study. Innovators were not found to be apinian leaders. However, there was some evidence that they increased their opinion leadership as a result of their innovation. This change may reflect changes in their social system, e.g., the norms that relate to innovation, and suggests that support of an innovator does have an influence on other faculty. The nature of this change in opinion leadership (such as under what conditions it occurs) and its effect on the department needs to be identified. 104 Replication with better controls. While this study accounted for 602 of the variance between innovators and non-innovators, it should be replicated with better controls. The actual adoption or nan—adoption of innovations should be documented. In this study, innovators were selected on the basis Pf reports of their innovation projects. However, the faculty were initially assigned as unsupported innovators or non-innovators on the basis of self-reports with no attempt to verify whether these reports were accurate. Stern et al. (1976) point out that attitudes toward adoption and actual adoption are not the same. The same problem may exist with self-reports. Replication is also needed to provide more precise measures of the variables. This could clear-up the problems with several of the variables which were described earlier in this chapter; these were problems which may have resulted in findings of no significant difference between the groups on those variables. More precise measures of the variables could also be used to establish the reliabil- ity of the self-reports used in this study. Such measures could take the form of confirmations by colleagues, use of records to verify self-reports, use of observers to record activities over a period of time, etc. The need for better controls also extends to the design of the study. This study used an ex post facto design--conducting the study after the significant event (innovation) had taken place. Furthermore, this study asked respondents in one group to base their ansers on different time periods than those in other groups. A more controlled design which takes these factors into account would provide a stronger foundation for findings which showed differences between innovators and non-innovators. 105 Summagy The purpose of this study was two-fold: first, it was to identify how innovators differed from non-innovators, and second, it was to identify how innovator's perceptions of their departments differed from those of non-innovators. The major findings of this study were that there were five characteristics which differed between innovators and non-innovators, and three characteristics which differed between the innovators' perceptions of their departments and those of the non-innovators. Furthermore, there were three distinct groups of faculty: innovators, early adapters, and non-innovators. A number of implications can be drawn from these findings. First, the differences found between the various groups of faculty can be used to identify whether an individual is more likely to be an innovator, an early adapter or a non-innovator. Second, strategies to increase the adoption of innovations can be based on making individuals more like innovators or early adapters in terms of the variables which differed between these two groups and the non- innovators. Third, EDP activities which are based an optional decisions to adapt innovations do not reach a wide cross section of faculty. Fourth, the adoption of innovations depends on the individual perceiving that there is adequate financial and psychological support for the innovation. Finally, a comprehensive theory is needed to explain how the variables in the Rogers and Shoemaker model on the diffusion and adaption of innovations fit together and interact with each other. BIBLIOGRAPHY BIBLIOGRAPHY Allen, T.M. Consequences when assumptions are not met. Unpublished manuscript, Michigan State University, no date. Anastasi, A. Psychological testing. New York: MacMillan Co., 1954. Berquist, W.H. & Phillips, S.R. Components of an effective faculty development program. Journal of Higher Education, March/April 1975, 46 (2), 177-211. Brickell, H. Organizing New York state for educational change. Albany, New York: New York State Education Department, 1961. Brown, D.G. & Hanger, W.S. Pragmatics of faculty self-development. Educational Record, Summer 1975,.26, 201-206. Campbell, D.T. & Stanley, J.C. Experimental and quasi-experimental desigps for research. Chicago: Rand MCNally, 1963. Carlson, R.O. Summary and critique of educational diffusion research. ,x’ In Michigan Department of Education, Research implications for educational diffusion. Lansing, Michigan: Michigan Department of Education, 1968. Cleland, S. Internships are second best. Liberal Education, October 1969, ii, 421-432. Crane, D. Invisible colleges--diffusion of knowledge in scientific communities. Chicago: University of Chicago Press, 1972. Davis, R.H. Personal and organizational variables related to the adoption of educational innovations in a liberal arts college (Doctoral dissertation, University of Chicago, 1965). (ERIC Document Reproduction Service No. ED 003 685) Davis, R.H., Abedor, A.J. & Witt, P.W.F. In_pursuit of excellence: a decade of instructional development at Michigan State Universipy. East Lansing, Michigan: Michigan State University, Educational Development Program, 1976. Demerath, J.J. & Daniels, L.A. How to make "the fourth revolution"-— human factors in the adoption of electronic instructional aids (Memorandum No. 73/5, Program on Application of Communication Satellites to Educational Development). St. Louis, Missouri: Washington University, 1973. (ERIC Document Reproduction Service No. ED 086 262) 106 Dietrich, J.E. & Johnson, F.C. A catalytic agent for innovation in higher education. Educational Record, Summer 1967, 48, 206-213. Evans, R.I. & Leppmann, P. Resistance to innovation in higher education San Francisco: Jossey-Bass, 1968. Gouldner, A.W. Cosmapolitans and Locals: toward an analysis of latent social roles. Administrative Science Quarterly, December 1957, _2, 281-306. Haines, M.F. The uses of knowledge as determinents of college curriculum. Paper presented at the annual meeting of the American Educational Research Association, Chicago, April 1974. (ERIC Document Reproduc- tion Service No. ED 087 374). Havelock, R.G. Planning for innovation through dissemination and utilization of knowledge. Ann Arbor, Michigan: University of Michigan, Social Research Bureau, 1969. James, L.M.L. A comparison of teaching effectiveness of tenured and nontenured faculty as measured by student evaluations (Doctoral dissertation, University of Florida, 1975). Dissertation Abstracts International, 1975, 36, 5082A. (University Microfilms No. 76-4242). Kazlow, C. & Giaquinta, J. Faculty_receppivity to organizational change. Paper presented at the annual meeting of the American Educational Research Association, Chicago, April 1974. (ERIC Document Repro- duction Service No. ED 090 847) Kerlinger, F.N. Foundations of behavioral research. New York: Holt, Rinehart & Winston, 1964. Kim, J. & Kohout, F.J. Multiple regression analysis: subpragram regression. Chap. in N. H. et al., Statistical Package for the Social Sciences (2nd ed.). New York: McGraw Hill, 1975. Klecka, W.R. Discriminant analysis. Chap. in H. H. Nie et al., Statistical Package for the Social Sciences (2nd ed.). New York: Mc-Graw Hill, 1975. Lewis, L.S. Two cultures: some empirical findings. Educational Record, Summer 1967, 48, 260—267. Mango, R.M. Faculty attitudes and teaching improvement. Improving College and University Teaching, Summer 1973, 21, 208-211. Michigan Department of Education. Research implications for educational diffusion. Lansing, Michigan: Michigan Department of Education, 1968. 107 Michigan State University, Educational Development Program. Educational Development at Michigan State University (6 Reports). East Lansing, Michigan: Michigan State University, Educational Development Program, 1971-1976. Miles, M.B. Innovation in education. New York: Columbia University, Teachers College, 1964. Mitchell, W.G. Communication of an educational innovation in an institution of higher learning (Doctoral dissertation, Michigan State University, 1970). Dissertation Abstracts International, 1970, 31, 5721A. (University Microfilms No. 7l-ll,928) Nie, N.H., Hull, C.H., Jenkins, J.G., Steinbrenner, K. & Brent, D.H. Statistical Package for the Social Sciences (2nd ed.). New York: MCGraw Hill, 1975. Peters, D.S. Faculty views of curriculum and instruction: a research report. Lubbock, Texas: Texas Technical University, Institute for the Improvement of Undergraduate Education, September, 1972. (ERIC Document Reporduction Service No. ED 077 367) Rogers, E.M. Diffusion of innovations. New York: The Free Press, 1962. ihfi Hm.pgi7. Rogers, E.M. The communication of innovations in a complex institution. Educational Record, Winter 1968, 42, 67-77. L7” Jeth- Ragers, E.M. & Jain, N.C. Needed research on diffusion within educational organizations. In Michigan Department of Education, ,, Research implications for educational diffusion. Lansing, Michigan: \/ Michigan Department of Education, 1968. Rogers, E.M. & Shoemaker, F.F. Communication of innovations-—a cross-cultural approach (2nd ed.). New York: The Free Press, ‘5”7/0’1357 1971. (Originally published as E.M. Rogers, Diffusion of innovations, 1962.) Rogers, E.M. & Svenning, L. Modernization among peasants——the impact of communication. New York: Holt, Rinehart & Winston, 1969. Sarbaugh, L.E. & Hawkins, B. An investigation of communication efforts and their relation to stages of adoppion of selected reading programs. East Lansing, Michigan: Michigan State University, Department of Communication, 1973. Stern, L.W., Craig, L.S., LaGreca, A.J. & Salem, R.G. Effect of sociometric location on the adoption of an innovation within a university faculty. Sociology of Education, January 1976, 42, 90-96. 108 Tatsuoka, M.M. Significance tests: univariate and multivariate. In M.M. Tatsuoka (Ed.), Selected topics in advanced statistics: an elementary approach (Vol. 4). Champaign, Illinois: Institute for Personality and Ability Testing, 1971. Winstead, P.C., Draine, D.P. & Romaine, B.H. The educational development officer--a catalyst for chapge in higher education (RELCV Working Paper No. 1-69). Durham, North Carolina: Regional Educational Laboratory for the Carolinas and Virginia, May 1969. 109 APPENDICES APPENDIX A MICHIGAN STATE UNIVERSITY Office of the Provost East Lansing - Michigan - 48824 APPENDIX A (Sent to EDP-Project Directors) In recognition of our 10th year of operation, we are writing a monograph describing the growth and development of the Educational Development Program (EDP) during the past decade. The monograph will focus primarily on the types of projects which have been supported and the growth of our instructional support program in response to the changing context at MSU. We would like to include a profile of characteristics of our EDP Project Directors and their departments. In order to construct this profile, we need your help. Please complete the enclosed questionnaire this week and return it in the enclosed, pre-addressed envelope by campus mail. We are working with an early deadline so a prompt response would be greatly appreci- ated. Our pilot tests have shown it will take you approximately 15-20 minutes to complete the questionnaire. I urge you to help us in this important endeavor. Thank you for your cooperation. Sincerely, Robert H. Davis, Assistant Provost Director, Instructional Development and Telecommunication Services ads Enclosure 110 111 EDP PROJECT DIRECTOR PROFILE NAME DEPARTMENT COLLEGE DATE OF FIRST EDP GRANT AMOUNT OF FIRST EDP GRANT INTRODUCTION This questionnaire has been designed to solicit information for our ten-year report on the MSU Educational Development Program. THERE ARE NO RIGHT ANSWERS and ALL RESPONSES WILL BE KEPT STRICTLY CONFIDENTIAL. The data will be reported in summary form only. Please put your completed questionnaire in the enclosed, pre-addressed envelope and return it by campus mail. We need to have all responses in our office no later than March 5th. The questionnaire deals with three time periods: The YEAR PRIOR to your EDP project, the TIME WHEN YOU BEGAN your EDP project, and the PRESENT TIME. Even though some of the questions ask you to think back a number of years, please try to do so. You will not need to look up or provide large amounts of factual data. Most of the items ask you to select a response which most closely represents how you feel about the item. lehich most closely describes what you believe to be the best answer. We know your time is limited, so the number of items has been limited. It should only take you 15-20 minutes to complete the questionnaire. Because this questionnaire is being sent to a very specialized audience, everyone's response is important. We would greatly appreciate it if you would take time to complete the question- naire within the next week. 112 ************************************************************************ 1. DURING THE YEAR PRIOR TO YOUR FIRST EDP GRANT, how often would you say you did the following? SCALE: 0=Never, l=0nce, 2=Severa1 (2-4 times), 3=Many (5 or more times) Attended a workshop, seminar, or meeting about teaching or learning at MSU. Attended a workshop, seminar, or meeting about teaching or learning at another university. Read a book or journal article about teaching or learning. Consulted with an MSU consultant on teaching, learning, evaluation, or media. Consulted with colleagues in your college about teaching or learning. Consulted with personal contacts at other universities about teaching or learning. Other: How did you first find out about the EDP program? (Check only your flgsp source of information.) A colleague Learning and Evaluation Service Instructional Media Center Instructional Television Service _____EDP literature or staff .____ Other (list) IN THE YEAR PRIOR TO YOUR FIRST EDP GRANT, how important was teach- ing compared to your other activities at MSU? (Mark your answer on the following scale.) / / / / / / / 1 2 3 4 5 6 7 Least important Equally important Most important 113 DURING THE YEAR PRIOR TO YOUR FIRST EDP GRANT, how important would you say good teaching was for promotion, pay raises, or other rewards in your department? / / / / / / / l 2 3 4 5 6 7 Not important Moderately important Very important DURING THE YEAR PRIOR TO YOUR FIRST EDP GRANT, approximately what percentage of your time did you spend directly preparing for teach- ing or actually teaching? Z PRIOR TO YOUR FIRST EDP GRANT, how much did you know about new developments in teaching and learning? / / / / / / / l 2 3 4 S 6 7 Knew very little Knew a great deal EDP project directors have given many reasons for seeking an EDP grant; how influential were each of the following in your decision to seek an EDP grant to improve your course? Please base your answers on the following scale: / / / / / / / l 2 3 4 5 6 7 Not influential Moderately Extremely I influential influential Your department chairman Your students Not satisfied with your own teaching Not satisfied with some aspect of the course For promotion, tenure, or pay raise For your personal development For recognition and status Increased enrollment Changes in the subject matter Changes in the curriculum 10. 11. 12. 114 DURING THE YEAR PRIOR TO YOUR FIRST EDP GRANT, how often would you say you participated in the activities of your department (meetings, seminars, social events, etc.)? / / / / / / / l 2 3 4 5 6 7 Rarely Half the time Almost always PRIOR TO YOUR FIRST EDP GRANT, which of the following teaching models were used in your department's courses? (CHECK ALL THAT YOU RECALL) Lecture Tutorial (one-to-one Recitation/seminar instruction) Laboratory/workshop Audio-tutorial.(self- Independent Study paced instruction using Field Study media) Practicum (Clearkship-Internship) Competency-based or mastery programs Other IN THE YEAR PRIOR TO YOUR FIRST EDP GRANT, how would you describe the human and material resources in your department available to help faculty members make changes to improve their instruction? / / / / / / / l 2 3 4 5 6 7 Extremely scarce Adequate More than sufficient PRIOR TO YOUR FIRST EDP GRANT, do you feel you were generally regarded by your colleagues as a good source of information or advice about teaching or learning? Yes No Compared to colleagues in your department PRIOR TO YOUR FIRST EDP GRANT, when were you most likely to adopt an educational innovation, e.g., programmed instruction, mastery learning, SLATEs, etc. (CHECK ONLY ONE)? Please base your answers to items 13 and 14 an the following scale: / / / / / / / 1 2 3 4 5 6 7 Hostile Neutral Enthusiastic 15. 115 13. DURING THE YEAR PRIOR TO YOUR FIRST EDP GRANT, when the tapic of new curriculum or new teaching methods was brought up, how would you describe the reaction of the faculty in your department? 14. How would you describe the reaction of the faculty in your department to your first EDP project? PRIOR TO YOUR EDP GRANT, how often did teaching assignments change in your department? / / / / / / / 1 2 3 4 5 6 7 Rarely changed Frequently changed (stayed the same (every quarter) for several years) *********************************************************************** ANSWERS IN THE NEXT SECTION SHOULD BE BASED ON THE TIME WHEN YOU BEGAN YOUR EDP PROJECT. PLEASE ANSWER THE FOLLOWING QUESTIONS BASED ON THE TIME WHEN YOU BEGAN THE EDP PROJECT. 9. Number of years you had been at Michigan State University Your faculty rank Were you tenured? Yes No Your age Your usual yearly teaching load, in student credit hours Number of different courses you taught each year Which course(s) were a part of your EDP project, e.g., ED882? Were you the only one in your department who taught this (these) course(s)? Yes No Approximately how many undergraduate students per year were affected by the EDP project? *********************************************************************** 116 PLEASE ANSWER THE REMAINING QUESTIONS BASED ON THE PRESENT TIME. 1. Approximately what percentage of your time do you spend directly preparing for teaching or actually teaching? 2. Do you feel you are generally regarded by your colleagues as a good source of information or advice about teaching or learning? Yes No 3. Have you been assigned additional teaching load(s) since your EDP project? Yes No 4. How much have each of the following influenced your teaching scale? Please base your answers on the following scale? / / / / / / 1 2 3 4 5 6 Not an influence Moderately Extremely influential influentia A former instructor of yours Colleagues in your department Colleagues at another university A consultant on the teaching/learning process (such as some- one from the Learning & Evaluation Service, IMC, ITV, etc.) A book or other publication Other (list) Please base your answers to items 5 and 6 on the following scale: / / / / / / / l 2 3 4 5 6 7 Strongly disagree Neutral Strongly agree] 5. Having a "good" class depends primarily on having bright students. 6. Some people are born to be good teachers; others are not. PEAK-LEN)” ”and, K ..r. L1 F.J.”u Fm ‘ . . in q 117 Is what you developed as a results of the EDP project still being used in your department? Yes No Are you using it? Yes No What changes occurred in the following as a result of your EDP project? SCALE: 1=Increased, 2=Decreased, 3=Stayed the same, 4=Don't know Enrollment Student learning Positive student ratings on SIRS or other end-of-term evaluations Please use the following word scale to answer the remaining items: 15. SCALE: 1=Definitely not, 2=Porbab1y not, 3=Probably yes, 4=Definitely yes 10. Do the faculty in your department consider your EDP project a success? 11. Do you consider your EDP project a success? 12. Was your EDP project worth the effort to you? 13. Would you consider another EDP project? 14. Since the completion of your EDP project, have you attempted any additional instructional innovation (with or without EDP funds)? Please add any comments you wish to make concerning your EDP project on the back of this form. THANK YOU for your help on this survey! APPENDIX B MICHIGAN STATE UNIVERSITY Office of the Provost East Lansing - Michigan - 48824 APPENDIX B (Sent to Non-EDP Supported Faculty) We are conducting a study on the growth and development of the Educa— tional Development Program (EDP) during the past decade. Part of this study will focus on a profile of selected characteristics of the entire MSU faculty. In order to construct this profile, your name has been randomly selected from the latest faculty list provided by the Provost's Office. We need your help by completing the question- naire. We know your time is limited so the number of items has been limited. Please take time to complete this questionnaire as soon as possible. Our pilot tests have shown it will only take about 15 minutes. THERE ARE NO RIGHT ANSWERS and ALL RESPONSES WILL BE KEPT STRICTLY CONFIDENTIAL. You will not need to look up or provide large amounts of factual data. Most of the items ask you to select a response which most closely describes what you believe to be the best answer. The data will be reported in summary form only and will be made available to you after the study is finished. Please put your completed questionnaire in the enclosed, self-addressed envelope. We need to have all responses in our office no later than April 30. IF YOU HAVE QUESTIONS CONCERNING THIS SURVEY, PLEASE CONTACT: Dr. Allan Abedor, Assistant Director of the Educational Development Program (353-1695). Thank you for your cooperation. Sincerely, Robert H. Davis, Assistant Provost Director, Instructional Development and Telecommunication Services mjs Enclosure 118 119 FACULTY SURVEY ON INSTRUCTION NAME DEPARTMENT COLLEGE l. DURING THE PAST YEAR, how often would you say you did the following: SCALE: O=Never, l=0nce, 2=Several (2—4 times), 3=Many (5 or more times) Attended a workshop, seminar, or meeting about teaching or learning sponsored by MSU. Attended a workshop, seminar, or meeting about teaching or learning sponsored by another university. Read a book or journal article about teaching or learning. Consulted with an MSU consultant on teaching, learning, eval- uation, or media. Consulted with colleagues in your college about teaching or learning. Consulted with personal contacts at other universities about teaching or learning. Other: DURING THE PAST YEAR, how important was teaching compared to your other activities at MSU? (Mark your answer on the following scale.) / / / / / / ‘ / l 2 3 4 5 6 7 Least important Equally important Most important DURING THE PAST YEAR, how important would you say good teaching was for promotion, pay raises, or other rewards in your department? (Mark your answer on the following scale.) / / / / / / / l 2 3 4 5 6 7 Not important Moderately important Very important 120 DURING THE PAST YEAR, approximately what percentage of your time did you spend directly preparing for teaching or actually teaching? AT THE PRESENT TIME, how much do you know about new developments in teaching and learning? / / / / / / / l 2 3 4 5 6 7 Know very little Know a great deal DURING THE PAST YEAR, how often would you say you participated in the activities of your department (meetings, seminars, social events, etc.)? / / / / / / / l 2 3 4 5 6 7 Rarely Half the time Almost always DURING THE PAST SEVERAL YEARS, which of the following teaching models were used in your department's courses? (CHECK ALL THAT YOU RECALL) Lecture Tutorial (one-to-one instruc- Recitation/seminar tion Laboratory/workshop Audio-tutorial (self-paced Independent study instruction using media) Practicum (Clerkship- Competency-based or mastery Internship) programs Other DURING THE PAST YEAR, how would you describe the human and material resources in your department available to help faculty members make changes to improve their instruction? / / / / / / / 1 2 3 4 5 6 7 Extremely scarce Adequate More than sufficient DURING THE PAST YEAR, do you feel you were generally regarded by your colleagues as a good source of information or advice about teaching or learning? Yes No 10. ll. 12. 13. 14. 15. 16. l7. l8. 19. 121 Compared to colleagues in your department, when are you most likely to adopt an educational innovation, e.g., programmed instruction, mastery learning, SLATEs, etc.? (CHECK ONLY ONE) Among the very first to try a new idea Before most others Just before the average faculty member Just after the average faculty member Among the last DURING THE PAST YEAR when the topic of new curriculum or new teaching methods was brought up, how would you describe the reaction of the faculty in your department? / / l / / / / 1 2 3 4 5 6 7 Hostile Neutral Enthusiastic DURING THE PAST FEW YEARS, how often did teaching assignments change in your department? / / / / / / / l 2 3 4 5 6 7 Rarely change Frequently change (stay the same (every quarter) for several years) Number of years you have been at Michigan State University: Your faculty rank: Are you tenured? ____Yes No Your age: Number of different courses you teach each year: Are you the only one in your department who teaches this (these) course(s)? Yes No How much have each of the following influenced your teaching style? Please base your answers on the scale below: / / / / / / / l 2 3 4 S 6 7 Not an Moderately Extremely influence influential influential 19. 22. 23. 122 A former instructor of yours Colleagues in your department Colleagues at another university A consultant on the teaching/learning process (such as someone from the Learning and Evaluation Service, IMC, ITV, etc.) A book or other publication Other (list) Please base your answers to Questions 20 and 21 on the scale below: / / / / / / / l 2 3 4 5 6 7 Strongly disagree Neutral Strongly agree 20. Having a "good" class depends primarily on having bright students. 21. Some people are born to be good teachers; others are not. In general, how do you currently teach? (Please mark each scale.) / / / / / / / l 2 3 4 5 6 7 I primarily lecture I rarely lecture / / / / / / / l 2 3 4 5 6 7 I use no audio- I use many audio- visual materials visual materials / / / / / / / l 2 3 4 5 6 7 All students move Students work at at the same pace-- their own pace-- must take exams at can take exams the same time when they're ready Compared to colleagues in your department, how would you describe your teaching techniques? / / / / / / / l 2 3 4 5 6 7 Almost identical About the same Very different 24. 123 DURING THE PAST FIVE YEARS, HAVE YOU BEEN INVOLVED IN AN INSTRUCTIONAL INNOVATION PROJECT (with or without outside funding)? Yes No If yes, please describe nature or goal of the project in a sentence or two: ***************************************it********************************* PLEASE BASE YOUR ANSWERS TO THE FINAL FIVE QUESTIONS ON YOUR MEMORY OF CONDITIONS FOUR OR FIVE YEARS AGO. 1. COMPARED TO FOUR 0R FIVE YEARS AGO, did you seek more or less information about teaching and learning during the past year? Sought more information Sought about the same amount Sought less information COMPARED TO FOUR OR FIVE YEARS AGO, is good teaching now more or less important for promotion, pay raises, or other rewards in your department? More important than before About the same as before Less important than before COMPARED TO FOUR OR FIVE YEARS AGO, have you participated more or less in the activities of your department (meetings, seminars, social events, etc.) during the past year? Participated more Participated about the same Participated less 124 4. COMPARED TO FOUR OR FIVE YEARS AGO, was the reaction of the faculty in your department more or less enthusiastic during the past year when the topic of new curriculum or teaching methods was brought up? More enthusiastic About the same Less enthusiastic 5. THINKING BACK FOUR OR FIVE YEARS, do you feel you were then generally regarded by your colleagues as a good source of infor- mation or advice about teaching or learning? Yes No THANK YOU for your help on this survey! APPENDIX C APPENDIX C SUMMARY DATA FROM EDP PROJECT DIRECTOR QUESTIONNAIRE RETURNS EDP PROJECT DIRECTOR PROFILE* *All data are percentages of the total faculty responding to an item. Totals may not add up to 1002 due to rounding. 1. Academic Discipline Natural Science 38 Social Science 29 Humanities 32 2. DURING THE YEAR PRIOR TO YOUR FIRST EDP GRANT, how often would you say you did the following? Never Attended a workshop, seminar, or meeting about teaching or learning at MSU. 36 Attended a workshop, seminar, or meeting about teaching or learning at another university. 54 Read a book or journal article about teaching or learning. 5 Consulted with an MSU consultant on teaching, learning, evalua- tion, or media. 45 Consulted with colleagues in your college about teaching or learning. 9 Consulted with personal contacts at other universities about teaching or learning. 46 Other: 125 Once 22 27 12 13 13 2-4 times 25 15 30 31 41 25 5+ times 16 54 10 47 15 3. How did you first find out (Check about the EDP program? only your first source of information). A colleague Learning and Evaluation Service Instructional Media Center Instructional Television Service EDP literature or staff Other (list) 4. IN THE YEAR PRIOR TO YOUR FIRST EDP GRANT, how important was teaching compared to your other activities at MSU? 5. DURING THE YEAR PRIOR TO YOUR FIRST EDP GRANT, how important would you say good teaching was for promotion, pay raises, or other rewards in your department? 6. DURING THE YEAR PRIOR TO YOUR FIRST EDP GRANT, approximately what per- centage of your time did you spend directly preparing for teaching or actually teaching? 7. PRIOR TO YOUR FIRST EDP GRANT, how much did you know about new develop- ments in teaching and learning? 126 48 27 14 Least Equally important important 1 2 3 4 5 0 6 10 25 18 Not Moderately important important 1 2 3 4 5 7 13 19 28 13 0-10Z ll-20Z Most important 6 7 15 27 Very important 6 7 12 7 21-30% 31-40% 41-50Z 3 3 ll 22 51-60% 61-70% 71-80% 81-902 91-100Z 15 9 15 6 Knew very Knew a little great deal .1. .2. .3. i 2 9 .7. 3 13 28 16 22 9 8 EDP project directors have given many reasons for seeking an EDP grant; how influential were each of the follow- ing in your decision to seek an EDP grant to improve your course? Your department chairman Your students Not satisfied with your own teaching Not satisfied with some aspect of the course For promotion, tenure, or pay raise For your personal development For recognition and status Increased enrollment Changes in the subject matter Changes in the curriculum DURING THE YEAR PRIOR TO YOUR FIRST EDP GRANT, how often would you say you participated in the activities of your department (meet- ings, seminars, social events, etc.)? 127 Not influential .1. .2. 48 8 27 13 25 4 15 0 68 9 18 2 49 10 47 15 53 4 49 10 Rarely l 2 0 3 Moderately influential .31 i .5. l6 6 6 10 22 6 10 18 7 3 13 10 10 10 3 7 13 27 13 15 10 2 10 3 7 7 15 4 10 9 Half the time 2 fl. 2 0 l3 l8 Extremely influentia 6 13 22 27 21 10 Almost alway 44 1 l .7. 10 9 13 32 l3 13 12 S 9 10. 11. 12. 13. 128 PRIOR TO YOUR FIRST EDP GRANT, which of the following teach- ing models were used TOTAL MODELS CHECKED in your department's courses? (Check all l_ .2 _3 that you recall used.) 2 15 19 ___Lecture ___Recitation/seminar __Laboratory/workshop __Independent study ‘__Field study __Practicum (Clerkship/Internship __Tutorial (one-to-one instruction) __Audio-tutorial (self-paced instruction using media) ___Competency-based or mastery programs __Other IN THE YEAR PRIOR TO YOUR FIRST EDP GRANT, how would you describe the human and material available to help Extremely faculty members make scarce changes to improve l_ .2 _3 their instruction? 18 37 18 PRIOR TO YOUR FIRST EDP GRANT, do you feel you were generally regarded by your colleagues as a good source of information or advice about teach- Ye§_ Np ing or learning? 49 52 Compared to colleagues in your department PRIOR TO YOUR FIRST EDP GRANT, when were you most likely to adopt an educational innovation, e.g., programmed instruction, mastery learning, SLATES, etc. (CHECK ONLY ONE)? i 18 Adequate i 16 2 18 .5. 8 6 16 More than sufficien 6 2 t l 2 13. 14. 15. 16. 129 (continued) Yes ‘Ng __Among the very first to try a new idea 47 __lBefore most others 39 __Just before the average faculty 11 __Just after the average faculty member 3 __Among the last 0 DURING THE YEAR PRIOR TO YOUR FIRST EDP GRANT when the topic of new curriculum or new teaching methods was brought up, how would you describe Hostile Neutral Enthusiastic the reaction of the ._l _l __3 _44 __5 ._6 _1 faculty in your 4 7 18 7 15 4 4 department? How would you des- cribe the reaction of the faculty in your department to Hostile Neutral Enthugiastic your first EDP _1. 3 3 __l: _E __ 1 project? 2 3 9 43 22 15 7 PRIOR TO YOUR EDP Rarely changed GRANT, how often (stayed the same Frequently changed did teaching for several years) (every quarter) assignments change 1 2 3 _4_ 5 6 in your department? 2T. 25- I5' 19 9 6 oq~J PLEASE ANSWER THE FOLLOWING QUESTIONS BASED ON THE TIME WHEN YOU BEGAN THE EDP PROJECT. 1. Number of years you had been at Michigan State 1-5 6-10 11-15 16-20 21-25 26-30 31-35 University: 36 20 18 14 9 3 2 Your faculty rank: ASST PROF ASSOC PROF PROF 32 24 If 44 Were you tenured? Your age: Your usual yearly teaching load, in student credit hours: Number of different courses you taught each year: Which course(s) were a part of your EDP project, e.g., ED882? Were you the only one in your department who taught this (these) course(s)? Approximately how many undergraduate students per year were affected by the EDP project? PLEASE ANSWER THE REMAIN- ING QUESTIONS BASED ON THE PRESENT TIME. 1. Approximately what percentage of your time do you spend directly preparing for teaching or actually teaching? Do you feel you are generally regarded by your colleagues as a good source of infor- mation or advice about teaching or learning? 130 Yes N9 69 31 Under 30 30-39 40-49 50-59 Over 60 15 29 31 19 6 OMMITTED FROM STUDY “I n20» oqox UH\J nfla> _. _‘1 9 16 34 19 OMMITTED FROM STUDY 28 72 TOTAL STUDENTS Under 100 101-200 201-300 301-400 31 22 ll 5 401-500 501-600 601-700 701-800 5 6 5 5 801-900 901-1000 Over 1000 3 3 8 O-lOZ ll-ZOZ 21-30% 31-40% 31-40% 51-60% 3 2 20 10 21 10 61-70% 71-802 81-9OZ 91-1002 15 10 7 3 Yes .N2 82 18 Have you been assigned additional teaching load(s) since your EDP project? How much have each of the following influenced your teaching style? A former instructor of yours Colleagues in your department Colleagues at another university A consultant on the teaching/learning process (such as someone from the Learning & Eval- uation Service, INC, ITV, etc.) A book or other publication Other Having a "good" class depends primarily on having bright students Some people are born to be good teachers; others are not. Is what you developed as a result of the EDP project still being used in your department? Are you using it? 131 Yes N9_ 28 71 Not an influential .1 .2 22 8 29 15 47 9 27 ll 26 12 74 2 Strongly disagree .51 .11 19 22 Strongly disagree 1. ._£ 20 14 Yes 1N3 78 22 Yes No 72 28 Moderately influential 2. fl. .2 6 19 15 14 9 12 9 ll 14 9 20 14 15 15 14 0 2 2 Neutral .il .11 .JE 26 15 15 Neutral .41 .41 ._§ 16 14 22 Extremely influential .9 .1 l7 l4 l4 6 8 3 12 8 9 12 Strongly agree .9 .Z 2 2 Strongly agree _2. .1 ll 3 132 What changes occured in the following as a result of your Increased Decreased Stayed the same Don't know EDP project? Enrollment 42 5 Student learning 70 2 Positive student ratings on SIRS or other end-of-term evaluations 41 2 Please use the following scale to answer the remain- Definitely Probably ing items: Not not 10. ll. 12. l3. 14. Teaching Specialization Do the faculty in your department consider your EDP project a success? 5 17 Do you consider your EDP project a success? 2 7 Was your EDP project worth the effort to you? 2 3 Would you consider another EDP project? 2 4 Since the completion of your EDP project, have you attempted any additional instructional innovation (with or without EDP funds? 6 4 COMBINED MEASURES 35 14 21 18 15 36 Probably Definitely yea-1 54 15 22 Not Specialized (No. of courses) 1 2 .3. 5*. .5. 8 9 22 13 11 Specialized (No. of courses) _1. _2. .3. i .5. 2 8 9 6 2 uflo~ hflO‘ 125: 25 77 91 72 84 WIN ON NHGD CMG) 133 Cosmopolite Information No. of Sources Used .2 .1 .l .2 _3. _5. _6 3 5 10 15 3 16 15 _7_ .§ _9 m .11 1; 1.7. 16 9 3 0 2 3 0 Local Information No. of Sources Used _9 _l _2. _2 __4. .2 .2 5 5 ll 21 18 24 9 Information Seeking No. of Sources Used .9 l .2. .3. .‘1 .5. .9 2 2 O 2 7 13 12 _7 _8 _.9_ 19 11 .11 .13. 4 4 9 13 9 l3 4 .11 1.5. 16. 1.7. 0 3 0 3 Cosmopolite Style Sum of Influences _9. _1 _Z _2. _4 _5_ .2 0 0 O 11 3 5 8 _z _8 .2 .19. 11 12. 1.3. 9 8 6 8 12 6 5 12 15. 16. 17. 1_8 12 11 3 2 2 2 3 Local Style Sum of Influences _0_ _l _2. _3. _4. .2 _6_ 2 0 9 6 5 O 12 _1 _§. .2 19 1.1. 1.2. 1;. 14 12 6 6 8 11 2 19. 1.5. 19 17. 1.8. 1.9. 2 O 0 0 0 0 Opinion Leadership Change Was an Opinion Leader (Before and after project Yes Before Not Before Yes Before Not Before or passage of time) Not After or After and After Yes After 0 18 52 31 Number of Projects .11 _;g _43 _43 -_5 ._6 ._1 ._§ 46 25 13 9 2 3 2 2 Combined Success (Possible Range 7-29) 134 5.: NO‘ IN O‘C: 26 10 27 13 APPENDIX D APPENDIX D SUMMARY DATA FROM NON-INNOVATORS QUESTIONNAIRE RETURNS FACULTY SURVEY ON INSTRUCTION* *All data are percentages of the total faculty responding to an item. Totals may not add up to 1002 due to rounding. 1. Academic Discipline Natural Science 43 Social Science 30 Humanities 27 2. DURING THE PAST YEAR how often would you say you did the following? Never Once 2-4 times Attended a workshop, seminar or meeting about teaching or learning at MSU 55 29 13 Attending a workshop, seminar, or meeting about teaching or learning at another university. 81 10 6 Read a book or journal article about teaching or learning. 12 12 36 Consulted with an MSU consultant on teaching, learning, evaluation or media. 71 15 9 Consulted with colleagues in your college about teaching or learning 12 8 33 Consulted with personal contacts at other universities about teaching or learning. 47 16 28 Other: . 93 1 1 135 5+ times 41 47 DURING THE PAST YEAR, how important was teaching compared to your other activities at MSU? DURING THE PAST YEAR, how important would you say good teaching was for promotion, pay raises, or other rewards in your department? DURING THE PAST YEAR, approximately what percentage of your time did you spend directly preparing for teaching or actually teaching? AT THE PRESENT TIME, how much did you know about new developments in teaching and learn- ing. DURING THE PAST YEAR, how often would you say you participated in the activities of your department (meetings, seminars, social events, etc.)? DURING THE PAST SEVERAL YEARS, which of the follow- ing teaching models were used in your department's courses? (Check all that you recall used.) __Lecture __Recitation/seminar __Laboratory/workshop __Independent study __Field study 136 Least Equally Most important important important _1_ .2 .2 .3. .2 .2 .1 l 6 26 29 25 13 Not Moderately Very important important important ..1_ .2 .2 .2 .2 .2 _7. 9 9 22 21 24 10 5 0-10% ll-ZOZ 21-30Z 31-4OZ 41-50Z O 3 13 15 20 51-60% 61-70% 71-80% 81-90 91-100% 16 12 18 3 0 Knew very Knew a little great deal ..1_ .2 .2 .51 .5. _6. _7. 13 13 22 28 17 6 1 Half Almost Rarely the time always _1_ _2 .3. __4 _5 .2 .1 2 4 2 12 18 23 39 TOTAL MODELS CHECKED _1. _Z. .2 .2 _5_ .2 .1+ 0 2 4 15 19 17 44 10. 11. 137 continued __Practicum (Clerkship/Internship) __JTutorial (one-to-one instruction) __lAudio-tutorial (self-paced instruction using media) __Competency-based or mastery programs __Other DURING THE PAST YEAR, how would you describe the human and material resources in your department available to help faculty Extremely members make changes scarce Adequate to improve their __l __2 __3 _44 __5 instruction? 11 23 16 31 ll DURING THE PAST YEAR, do you feel you were generally regarded by your colleagues as a good source of infor- mation or advice about :22. No teaching or learning? 71 29 Compared to colleagues in your department, WHEN are you most likely to adopt an educational innova- tion, e.g., programmed instruction, mastery learning, SLATES, etc. (CHECK ONLY ONE)? Among the very first to try a new idea 10 Before most others 30 Just before the average faculty 40 Just after the average faculty member 9 __ Among the last 12 Mare than sufficient 6 6 7 2 12. 13. 14. 15. 16. 17. 18. 19. 20. 138 DURING THE PAST YEAR, when the topic of new curriculum or new teaching methods was brought up, how would you describe the reaction of the Hostile faculty in your ._l _Jl department? 1 5 DURING THE PAST FEW YEARS, ing assignments change in your 1 2 department? 7 '23 Number of years you have been at Michigan State Neutral 4 5 _E' 2E' 40 U) Rarely changed how often did teach- (stayed the same for several years) .2 .22 .2. 12 22 16 Enthusiastic .2. ..1 15 2 .2. 14 Frequently changed (every quarter) ..1 7 1-5 6-10 11-15 16-20 21-25 26-30 31-35 University: 22 25 23 12 7 9 Your faculty rank: ASST PROF ASSOC PROF PROF 25 25 50 Are you tenured? Yes ‘52 81 19 Your age: Under 30 30-39 40-49 50-59 Over 60 3 34 27 18 Number of different courses you teach __l _Jl _ll _Jg _Ji _Ji __1 each year: 5 18 23 19 14 10 2 Are you the only one in your department who taught this Yes _Np (these) course(s)? 33 67 How much have each Not an Moderately Extremely of the following influential influential influential influenced your _l _2 _3_ __4_ _5 _6_ __1 teaching style? A former instruc- tor of yours 12 6 2 23 16 25 16 Colleagues in your department l8 l9 16 26 ll 6 5 .oloo 21. 22. 23. Colleagues at another university A consultant on the teaching/learning process (such as someone from the learning and Eval- uation Service, IMC, ITV, etc.) A book or other publication Other Having a "good" class depends primarily on having bright students. Some people are born to be good teachers; others are not. In general, how do you currently teach? (Please mark each scale) 139 38 16 71 10 38 16 80 Strongly disagree _1.._2 17 15 Strongly disagree _1...2. l7 14 I primarily lecture 22 24 24 I use no audio-visual materials .2_?. 12 15 All students at the same pace--must take 18 24 11 16 Neutral __4_ 35 Neutral .2 19 nflc~ H NI» exams at the same time 1 _2. 46 31 soLo mls aqua 7 O 0 1 5 2 8 9 Strongly agree .2 .1 7 3 Strongly agree .2 __7 14 5 I rarely lecture 2 1 6 0 I use no audio-visual materials .2 .1 22 10 Students work their own pace-- .2 7 can take exam when ready _6_ 3 .1 o 24. Compared to colleagues in your department, how would you des- cribe your teaching techniques? 25. DURING THE PAST FIVE YEARS, HAVE YOU BEEN INVOLVED IN AN INSTRUC- TIONAL DEVELOPMENT OR INSTRUCTIONAL INNOVA- TION PROJECT (with or without outside fund- ing?) 140 If yes, please describe nature or goal of the project. PLEASE BASE YOUR.ANSWERS TO THE FINAL FIVE QUESTIONS ON YOUR MEMORY OF CONDITIONS 0R FIVE YEARS AGO. l. COMPARED TO FOUR OR FIVE YEARS AGO, did you seek more or less information about teaching and learning during the past year? 2. COMPARED TO FOUR OR FIVE YEARS AGO, is good teaching more or less important for promotion, pay raises, or other rewards in your department? 3. COMPARED TO FOUR OR FIVE YEARS AGO, have you participated more or less in the activities of your department (meetings, seminars, social events, etc.) dur- ing the past year? 38 '17." 3 Almost About identical the same 2.2.2.4. 0 0 3 41 1921.9 0 88 Sought Sought more about the information same 24 64 More About important the same than before as before 25 64 Partici- Partici- pated more pated the same 15 77 Very different 5- 6 7 Sought less information 13 Less important than before 11 Partici- pated less 8 141 4. COMPARED TO FOUR OR FIVE YEARS AGO, was the reaction of the faculty in your department more or less enthusiastic during the past year More Less when the topic of enthusi- About the enthusi- new curriculum or astic same astic teaching methods 14 72 14 was brought up ? 5. THINKING BACK FOUR 0R FIVE YEARS, do you feel you were then generally re- garded by your col- leagues as a good source of informa- tion or advice about Yes. .EQ teaching or learning? 63 37 COMBINED MEASURES Teaching Specialization Not Specialized (No. of courses) _1.2._3.i_2_6_7 0 9 l6 16 10 3 2 .22 6 2 Specialized (No. or courses) 2.2.2.2221 5 9 7 3 2 6 0 Cosmopolite Information No. of Sources Used _Q_1_2.2_é._.2.2 8 3 18 18 18 14 6 2219111212 5 6 0 0 0 0 Local Information No. of Sources Used _Q_1.2_3..‘1..2.2 7 7 26 29 21 9 l I-aloo Lulu 142 Information Seeking No. of Sources Used .9. .1. .2 .2 .2 _5 .2.1 5 1 3 8 10 9 13 15 .8. .2 .12 11 12. .12 .1212 10 9 3 3 7 12 l 0 19.11 0 0 Cosmopolite Style Sum of Influence .9 .1. .2 .2 .2 .2 .2.1 0 0 0 22 11 8 l3 8 .2 .9. m 11 12. .12 1711.5. 7 10 2 6 4 2 5 0 191.7. 12.1.9. 1 0 O 0 Local Style Sum of Influence _0 ..1. .2 .2 .2 .2 .2_7 0 O 7 2 5 11 7 6 .2 _9 .12 u 12 .12 11.12 11 19 13 ll 7 0 l 0 12.11 1.812 0 0 0 0 Opinion Leadership Was an Opinion Leader Change Yes Before Not Before Yes Before Not Before (Before and after project Not after or After and After Yes After or passage of time) 4 26 60 10 APPENDIX E APPENDIX E SUMMARY DATA FROM UNSUPPORTED INNOVATORS QUESTIONNAIRE RETURNS FACULTY SURVEY ON INSTRUCTION* *All data are percentages of the total faculty responding to an item. Totals may not add up to 1001 due to rounding. 1. Academic Discipline Natural Science 44 Social Science 29 Humanities 27 2. DURING THE PAST YEAR how often would you say you did the following? Never Once 2-4 times 5+ times Attended a workshop, seminar, or meeting about teaching or learning at MSU. 42 22 24 12 Attended a workshop, seminar, or meeting about teaching or learning at another university. 68 22 6 5 Read a book or journal article about teaching or learning. 7 11 36 46 Consulted with an MSU consultant on teach— ing, learning, evalua- tion, or media. 45 21 21 13 Consulted with colleagues in your college about teaching or learning. 5 5 22 68 Consulted with personal contacts at other univer- sities about teaching or learning. 30 23 28 19 Other: 92 5 O 4 143 DURING THE PAST YEAR, how import- ant was teaching compared to your other activities at MSU? DURING THE PAST YEAR, how import- ant would you say good teaching was for promotion, pay raises, or other rewards in your department? DURING THE PAST YEAR, approximately what percentage of your time did you spend directly preparing for teach- ing or actually teaching? AT THE PRESENT TIME, how much did you know about new developments in teaching and learn- ing? DURING THE PAST YEAR, how often would you say you participated in the activities of your department (meet- ings, seminars, social events, etc.?) DURING THE PAST SEVERAL YEARS, which of the following teaching models were used in your depart- ment's courses? (Check all that you recall used.) __Lecture __Recitation/seminar __4Laboratory/worksh0p __lndependent study 144 Least important .1.1 O 1 Not important .13. 6 15 old 3 16 Equally important 4 .2 7 27 Moderately important _4..9. 27 16 Most important 1.1 16 23 Very important 1.7. 12 7 0-102 ll-ZOZ 21-30Z 31-40% 41-50% 2 2 11 10 21 51-60% 61-70% 71-80Z 81-9OZ 91-100Z l7 Rarely TOTAL MODELS CHECKED HI... .2. 1 9 "l o~w _3 0 old 18 7 .9; 35 19 Half the time _3..5. 9 15 col» .9. 8 4 Knew a great deal .9 _7 7 4 Almost always .9. .7. 43 29 .9. .7.+ 17 58 9. 10. 11. 145 ___Field study ._;Practicum (ClerkshOp/Internship) __Tutorial (one-to—one instruction) __Audio-tutorial (self-paced instruction using media) .__Competency-based or mastery programs __Other DURING THE PAST YEAR, how would you des- cribe the human and material resources in your department available to help faculty members Extremely make changes to scarce Adequate improve their _1_ ..§ __3 ._4 ._§ instruction? 7 19 24 23 18 DURING THE PAST YEAR, do you feel you were generally regarded by your colleagues as a good source of information or ad- vice about teaching ‘Ygg No or learning? 83 17 Compared to colleagues in your department, WHEN are you most likely to adopt an educational innova- tion, e.g., pro— grammed instruction, mastery learning, SLATEs , etc. (CHECK ONLY ONE)? Among the very first to try a new idea 23 Before most others 52 Just before the average faculty member 18 Just after the average faculty member 5 Among the last 3 More than sufficient 6 8 7 0 146 DURING THE PAST YEAR, when the tapic of new curriculum or new teaching methods was brought up, how would you 13. 14. 15. 16. 17. 18. 19. 20. describe the reac- Hostile Neutral Enthusiastic tion of the faculty __1_ __2 __3 _4 _5_ _6_ _Z in your department? 0 6 10 34 33 16 l DURING THE PAST YEAR, how often did teaching assignments change in your department? Number of years you have been at Michigan State Rarely changed (stayed the same for several years) 1 2 3 6 T5 '17 4 15 Frequently changed (every quarter) .2 28 6 7 7 11 1-5 6-10 11-15 16-20 21-25 26-30 31-35 University: 25 36 14 10 7 5 4 Your faculty rank: ASST PROF ASSOC PROF PROF 26 29 45 Are you tenured? Yes ‘N2 79 21 Your age: Under 30 30-39 40-49 50-59 Over 60 O 37 3O 23 11 Number of different courses you teach “_I ._2 ._3 ._4 __5 ._6 ._1 each year: 6 10 30 24 14 10 4 Are you the only one in your department who taught this Yes 'Ng (these) course(s)? 26 74 How much have each of the following influ- Not an Moderately Extremely enced your teach- influential influential influential ingscyle? .1. .2. _3. .1 _5_ .1 .1 A former instructor of yours 12 10 5 19 l7 l9 l9 Colleagues in your department 10 15 14 28 19 12 3 Colleagues at another university 29 23 11 23 3 10 3 blm 21. 22. 23. 147 A consultant on the teaching/learning process (such as someone from the Learning & Evaluation Service, IMC, ITV, etc.) 47 14 9 A book or other publi- cation 25 15 16 Other 72 O 0 Having a "good" class Strongly depends primarily disagree on having bright __1 “_2 .43 students. 12 20 13 Some people are Strongly born to be good disagree teachers; others are ._1 ._2 __3 not. 15 1 13 I primarily lecture In general, how do you currently teach? (Please “_1 ._2 ._3 mark each scale) 13 18 23 I use no audio- visual materials ..1 ._2 ._2 4 7 9 All students at the same pace-— must take exams at same time 1 2 ._2 44 18 14 16 26 Neutral ..2 28 Neutral .42 23 ”I 4‘“ ”I H‘b- Hl h‘¢~ 6 4 5 10 6 3 7 9 ll Strongly agree ._2 ._9 ..Z 17 4 6 Strongly agree .2. .9. .1. l7 l3 6 I rarely lecture ._2 .9. .1. 16 5 3 I use many audio- visual materials 6 ..2 ..Z 22 20 17 Students work their own pace- can take exams when ready oolm .51 4 HIV 24. Compared to colleagues in your department, how would you describe your teaching techniques. 25. DURING THE PAST FIVE YEARS, HAVE YOU BEEN INVOLVED IN AN INSTRUC- TIONAL DEVELOPMENT OR INSTRUCTIONAL INNOVAr TION PROJECT (with or without outside fund- ing?) If yes, please describe nature or goal of the project. PLEASE BASE YOUR ANSWERS TO THE FINAL FIVE QUESTIONS ON YOUR MEMORY OF CONDITIONS FOUR OR FIVE YEARS AGO. l. COMPARED TO FOUR OR FIVE AGO, did you seek more or less information about teaching and learning during the past year? 2. COMPARED TO FOUR OR FIVE YEARS AGO, is good teach- ing more or less import- ant for promotion, pay raises, or other rewards in your department? 3. COMPARED TO FOUR OR FIVE YEARS AGO, have you participated more or less in the activities of your department 148 Almost identical 2 3 Yes No Sought more information 32 More important than before About the same a .5 l7 4 \l Sought about the information 24 Participated more (meetings, seminars, social events, etc.) during the past year? 18 49 About the same as before 58 Participated the same 75 Very different __9. .7. 7 4 Sought less information 20 Less important than before 19 Participated less 8 4. COMPARED TO FOUR OR FIVE YEARS AGO, was the reaction of the faculty in your department more or less enthusiastic during the past year when the topic of new curriculum or teaching methods was brought up? 5. THINKING BACK FOUR OR FIVE YEARS, do you feel you were then generally regarded by your colleagues as a good source of information or advice about teach- ing Or learning? Teaching Specialization Cosmopolite Information Local Information More 149 enthusiastic Yes 25 11.0. 37 About the same 61 COMBINED MEASURES NIH HICD 1|... g...» Lulu blo Nlo ml... IQLO m'w NI... oxioo NI... Not Specialized (No. of courses) _3_2_2 21 17 13 Specialized (No. of courses) .2 8 .2 7 _5 1 No. of Sources Used SIN ml... .2 12 11 No. of Sources Used 2 l3 .2 35 _4 17 Less enthusiastic l4 .2_7 8 4 _9_l l 0 .2.9. 17 10 .1111 1 1 .2_9 18 12 0'00 150 Information Seeking No. of Sources Used 2.1.1.1222 l l 1 2 4 7 l6 22.211.11.22 8 7 12 7 ll 7 10 1115.16.11 1 2 l l Cosmopolite Sum of Influence .0_1_1_1_1_1.6. O O O 6 7 5 10 2221111111.?» 15 16 5 10 5 6 5 1121111112 4 5 1 O O 0 Local Style Sum of Influence 2.1.1.3122 O O 4 4 5 4 5 22219111113. 7 17 22 17 6 5 3 215.16.211.11 1 O O O O 0 Opinion Leadership Was an Opinion Leader Change (Before and after Yes Before Not Before Yes Before Not Before project or passage Not After or after and after Yes After of time) 3 15 61 21 MICHIGAN STATE UNIV. LIBRRRIES 1|1ll”1111111111111111111111111111111111 31293101155756