MSU LIBRARIES RETURNING MATERIALS: P1ace in book drop to remove this checkout from your record. FINES wi11 be charged if book is returned after the date stamped below. #26534? , EG'BBQ/ // THE ADOPTION AND APPLICATION OF COMPUTER-BASED TELECOMMUNICATION TECHNOLOGIES BY HOME CONSUMERS By Richard Vincent Ducey, Jr. A DISSERTATION Submitted to Hichigan State University in partiai fulfiilment of the requirements for the degree of DOCTOR OF PHILOSOPHY Coiiege of Communication Arts and Sciences Mass Media 1983 Copyright by RICHARD VINCENT DUCEY, JR. 1983 ABSTRACT THE ADOPTION AND APPLICATION OF COMPUTER-BASED TELECOMMUNICATION TECHNOLOGIES BY HOME CONSUMERS By Richard Vincent Ducey, Jr. Recent developments in technology, policy and industry standard- ization have stimulated growth in the number of possible applications of information processing technology. The ponderous question of whether better information processing and telecommunication machines can be built is no longer relevant. The issue to be recognized now is, given the multitude of design alternatives,which options should the designers of home information systems select? On what basis should these decisions be made? How can human satisfaction and competence with these sophisticated systems be maximized? These are the issues contemplated in the present research effort. The underlying theory is that given a set of communication needs, people will try to satisfy these needs by seeking out and adopting various technologies. Those technologies which embody the most salient attri- butes should be adopted more successfully. To study the process by which home consumers adopt telecommunicae tion products and services and apply these to serving communication needs, a two study approach was employed. The first study was designed to collect information from home consumers regarding their self- perceived communication needs and the degree to which various Ducey attributes of telecommunication technologies were salient to them. The second study used a_Delphi technique administered to a sample of telecommunication experts knowledgeable about technology. In this study. a set of estimates was derived indicating the extent to which various telecommunication technologies were appropriate for providing a number of computer-based services. Major findings of the consumer study are that human communication needs are significantly related to the salience of technology attri- butes. The importance of technology attributes to consumers predicts their economic value. Present telephone usage patterns also predicted the economic value of technology attributes. Several intangible attri- butes such as attitudes towards computerization and social interaction can explain some variation in the perceived salience of technology attributes. Finally, the Delphi Study revealed that two-way cable and telephone technologies were the most appropriate media for the provision of computer-based services. Limitations of the research and future directions are discussed. ACKNOWLEDGEMENTS I owe my largest debt of gratitude to my dissertation advisor, John D. Abel. His interest in and support of my work helped to both encourage and enlighten me as I completed this project. The other members of my committee have also influenced my thinking and ultimately the direction of this work. I express my deeply felt appreciation to Thomas F. Baldwin, Martin P. Block, John B. Eulenberg and Thomas A. Muth for serving on my committee. I must also thank Dean M. Krugman for those many brainstorming sessions, which in part lead to this disserta- tion. The completion of this dissertation is understandably an important accomplishment in my life. I am very grateful to the fine Mass Media Ph.D. program at Michigan State for providing me with the training and experience necessary to achieve this goal. I have only good things to say about this program. I would like to thank the Department of Telecommunication at Michigan State for their support in helping me to complete this work. Ann Alchin worked with me through various drafts as much more than a typist. She detected many subtle flaws in the manuscript which escaped my notice and I appreciate her suggestions and attention to detail. The students of my Audience Survey and Analysis class collected the data for the consumer study. They did a fine job as survey researchers. Another person who took great personal interest in the success of this project and contributed beyond the call of duty was Neal S. Yonover. My wife, Sara, is perhaps my most influential colleague. She has a very special way of listening to my babblings and somehow reflecting these back to me as polished ideas. Her keen insight and ability to articulate things has always helped me to develop my ideas into some- thing worthwhile. I have always felt that there is something magical about her. Finally, I acknowledge the love and support that my two sets of parents have shown. Sara's parents and my parents contributed much to our creature comforts and human nourishment. Their concern and generousity allowed me to pursue more lofty goals. I hope that I can live up to their expectations and the example they set. TABLE OF CONTENTS Page LIST OF TABLES ......................... vii LIST OF FIGURES ......................... x LIST OF APPENDICES ....................... xi CHAPTER I. INTRODUCTION .................... 1 Organization of report ................... 2 Problem statement ..................... 3 Review of the literature .................. 7 Attributes of innovations ................. l7 Perception of innovations ................. 18 Adoption process ...................... 2l Summary .......................... 29 Reference notes ...................... 3l CHAPTER II. RESEARCH APPROACH ................. 35 Conceptual overview .................... 35 Rationale for conducting two studies . ........... 4l Research hypotheses for the consumer study ......... 45 -— Research questions for the Delphi study .......... 48 Conclusion ......................... 49 Reference notes ...................... 50 CHAPTER III. METHODS FOR THE HOME CONSUMER STUDY ........ 52 Research design ...................... 52 Selection of variables ................... 52 Selection of the sample .................. 59 Administration of the survey ................ 60 Sample demographics .................... 6l Data reduction ....................... 64 Results of factor analyses ................. 71 Analytical techniques used in hypothesis-testing ...... 76 Reference notes ...................... 83 CHAPTER IV. HOME CONSUMER STUDY: RESULTS OF HYPOTHESIS TESTING. 88 Hypothesis l ........................ 88 Hypothesis 2 ........................ 92 Hypothesis 3 ........................ 93 TABLE OF CONTENTS (continued) Page CHAPTER IV (continued) Hypothesis 4 ........................ 95 Hypothesis 5 ........................ 96 Hypothesis 6 ........................ 97 Hypothesis 7 ........................ 98 Hypothesis 8 ........................ 100 Hypothesis 9 ........................ 100 Hypothesis lO ....................... 101 Discussion of results ................... 104 Reference notes ...................... 107 CHAPTER V. METHODS FOR THE TELECOMMUNICATION EXPERTS STUDY . . . 108 Research design ...................... 108 Instrumentation ...................... ll3 Selection of the sample .................. 116 Administration of the Delphi survey ............ 119 Sample demographics .................... 121 Reference notes ...................... 125 CHAPTER VI. RESULTS OF THE TELECOMMUNICATION EXPERTS STUDY . . . 127 One-way cable television .................. 129 Two-way cable television .................. 131 » Broadcast television .................... 131 Broadcast radio ...................... 133 Telephone ......................... 136 Direct broadcast satellite ................. 136 Multipoint distribution service .............. 138 Low power television .................... 138 Terminal design ...................... 138 addressability .................... l4l microprocessor .................... 143 keyboard type ..................... 143 memory ........................ 144 Discussion of results ................... 144 Reference notes ...................... 150 CHAPTER VII. SUMMARY AND DISCUSSION .............. 152 Diagnostic methods and reinvention ............. 152 Tangible and intangible attributes ............. 161 ”*“‘ Present attitudes and future adoption ........... 163 ‘ Delphi technique ...................... 164 «« Integrating major findings of Consumer and Delphi studies . 165 Limitations of this research ................ 167 Future directions ..................... 169 Conclusions ........................ 170 Reference notes ...................... 172 APPENDICES ........................... 174 vi TABLE II. III. IV. VI. VII. VIII. IX. XI. XII. XIII. XIV. XV. LIST OF TABLES Variables Used to Operationalize Needs, Attributes and Technologies Measurement Spaces ............ Salient Characteristics of Computer-Based Telecommunica- tion Systems Identified in the Research Literature Distribution of Sample Survey Telephone Calls ..... Sample Demograhics .................. Innovative and Non-Innovative Communication Products and Services ..................... Cronbach's Alpha Coefficient Reliability Estimates for Needs, Services and Features Scale ........ Eigenvalues and Percent of Variance Explained by Factors in Unrotated and Varimax Rotated Factor Analyses of Communication Needs ............ Varimax Rotated Factor Loading Matrix of Communication Needs ......................... Factor Score Coefficient Matrix of Communication Needs. Eigenvalues and‘ Percent of Variance Expalined in Unrotated and Varimax Rotated Factor Analyses of Saliency of Computer-Based Services .......... Varimax Rotated Factor Loading Matrix of Saliency of Computer-Based Services ................ Factor Score Coefficient Matrix of Saliency of Computer- based Services .................... Eigenvalue and Percent of Variance Explained by Factors in Unrotated Factor Analysis of Saliency of Features. . Canonical Correlates of Communication Needs and Saliency of Computer-Based Services Canonical Loading Matrix of Communication Needs and Saliency of Computer-Based Services ...... . . . . vii PAGE 42 56 62 63 65 68 72 73 75 76 77 78 79 89 91 TABLE XVI. XVII. XVIII. XIX. XX. XXI. XXII. XXIII. XXIV. XXV. XXVI. XXVII. XXVIII. XXIX. XXX. XXXI. XXXII. Multiple Regression of Number of Communication Products and Services Adopted on Communication Needs ...... Multiple Regression of Number of Innovations Adopted on Communication Needs ................ Multiple Regression of Amount Willing to Pay For Services Package on Communication Needs ........ Multiple Regression of Number of Communication Products and Services Adopted on Media Use ........... Multiple Regression of Number of Innovations Adopted on Media Use ..................... Multiple Regression of Amount Willing to Pay for Services Package on Media Use ............. Multiple Regression of Amount Willing to Pay fOr Services Package on Features Scale .......... Multiple Regression of Amount Willing to Pay for Services Package on Importance of Services ...... Pearson Product-Moment Correlation of Intangible Attributes With Services and Features Scales ..... Delphi Respondents' Self Rating on Knowledge of Telecommunication Technologies ............ Sample Demograhics for Telecommunication Experts in Rounds I and II of the Delphi Study .......... Delphi Ratings of Services and Appropriate Technologies: One-Way Cable Television Delphi Ratings of Services and Appropriate Technologies: Two-Way Cable Television ............... Delphi Ratings of Services and Appropriate Technologies: Broadcast Television .......... Delphi Ratings of Services and Appropriate Technologies: Broadcast Radio ............ Delphi Ratings of Services and Appropriate Technologies: Telephone ............... Delphi Ratings of Services and Appropriate Technoligies: Direct Broadcast Satellites viii Page 93 94 95 97 98 99 100 101 102 119 122 130 132 134 135 137 139 TABLE XXXIII. XXXIV. XXXV. Delphi Ratings of Services and Appropriate Technologies: Multipoint Distribution Service . . . . 140 Delphi Ratings of Services and Appropriate Technologies: Low Power Television ......... 142 Selection of Terminal Design Attributes by Round . . . 145 ix LIST OF FIGURES Page Figure 1. Conceptual Model of the Adoption Process ...... 36 Figure 2. Model of the Innovation/Adoption Process ...... 154 LIST OF APPENDICES Page Appendix A. Telephone Survey Instrument for Consumer Study . . . 174 Appendix B. Instrumentation for Delphi Study - Wave I ..... 182 Appendix C. Instrumentation for Delphi Study - Wave II ..... 189 xi CHAPTER I INTRODUCTION Driven by imaginative applications of computer-based and telecom- munication technologies, the dawning of the information age is rapidly becoming a personal reality for home consumers. Sophisticated informa- tion and communication technologies have been readily accessible to business and institutional users for years. The size, expense and level of sophistication required for successful operation has tended to inhibit the growth of these products and services into the consumer market. Small, affordable and user-friendly consumer oriented versions of these technologies now loom on the horizon. In fact, the rapid proliferation of small home computers and video games seems indicative of the interest and energy home consumers find for these technologies. Much research has been directed toward an exploration of the relationships which exist or will probably exist between the human users of these systems and the affinities of the technologies themselves. There is an interest in designing computer-based systems so that they are efficient from a human viewpoint rather than emphasizing machine efficiency. This research has considered such problems as how the systems should be designed; who will use them; what are their skills; what are their needs; who will spend money to gain access to these systems? The success of these systems depend on the quality of the human-machine interface. While humans can adapt themselves to situa- tions, efficiency, productivity and satisfaction are enhanced when 1 systems are instead adapted to human users. It was in this spirit that the present research was conducted. Organization of report Two separate studies were conducted in this research project. The first study surveyed a random sample of East Lansing residents. The second study used a Delphi technique to survey a purposive sample of respondents knowledgeable about telecommunication technologies. This chapter introduces the research problem and the significance of this research effort. To establish this conceptual framework, a review of the relevant research literature is presented. Chapter II considers the rationale and underlying objectives for the two studies. The study of East Lansing home consumers had the major objective of testing several hypothetical relationships regard- ing their communication needs, attitudes toward innovation attributes and the adoption and application of innovations. The purpose of the Delphi study was to estimate the extent to which a number of telecom- munication technologies are appr0priate for the provision of various telecommunication services and features to home consumers. The research questions addressed in this study are also discussed in this chapter. Chapter III focuses on the methods used in the home consumer study. Chapter IV relates the data analytic procedures used and the statistical testing of the hypotheses for this study. Chapter V is concerned with the methods of the Delphi study. Chapter VI presents and interprets the descriptive findings of this study. 3 Finally, Chapter VII summarizes the research findings and draws tentative conclusions based on these empirical findings. Directions for new research in this area are proposed. Problem Statement An interesting occurrence is the breakdown in the familiar boundaries among telecommunication marketplaces. Services no longer need to be associated with particular administrative or technological entities in telecommunication. In fact, similar services can be pro- vided by technologically distinct telecommunication entities. Tele- phone companies are seeking to provide competition to cable television companies for the provision of broadband services. Paging services are offered by both broadcasters and common carriers. High speed data communications are available by satellite or several terrestrial alternatives. Videotex services can be provided via telephone, tele- vision, cable television, or broadcast radio. Although similar video- tex services can be provided by different media, there are some differences. The type of service provided over these facilities will vary according to the technical affinities of the medium and other system variables. Thus, the telephone may permit two-way communica- tion; FM radio may be cheaper and television may be faster but be limited in the size of its database. The trend towards a redefinition of the telecommunication market- place in terms of emerging innovations seems clear. These "innova- tions" are sometimes entirely new technologies, such as direct broadcast satellites or cellular radio and sometimes new applications of existing technologies such as imaginative uses of FM radio subchannels. To a large extent, many of these new developments are 4 being driven by advancements in telecommunication applications of computers. A number of computer-based telecommunication technologies are in various stages of reaching the home consumer market. Some technologies such as two-way videotex are likely to be first targeted toward the business market before being rolled-out in the consumer market. Some testing has been going on around the nation to gauge preliminary interest in this type of interactive service. Likewise, one-way teletext systems have been introduced on a trial basis in several areas. Other technologies such as home computers, cable television, and various forms of pay television have already made a substantial impact on the consumer market. The "impact" of technology on the consumer market can be approached in several ways. One way is to consider the impact on consumption patterns. In some cases, market segments will be attracted away from their current consumption patterns and enticed into bringing old habits to new media. In other cases, entirely new appetites for services may be whetted, leading to the instigation of new markets. Consumption effects in the former case are known as "functional dis- placement." In this case, the new media do a superior job of meeting needs at an equal or smaller cost to the consumer. In the latter case, a form of "functional placement" occurs. In this case, no consumption pattern exists prior to the introduction of the technology. After the technology is introduced, new needs are "discovered" which only the new technology can serve. An example might be the introduction of real time polling capability with interactive cable television which provides the home viewer with the means to respond to multiple choice questions. In a complex environment of technological change, it will be difficult to predict which innovations will appeal to consumers and thus survive in the marketplace. Functional displacement effects are much easier to predict than functional placement effects. This is true because of the relative difficulty of trying to predict the emergence of "new" needs. In some instances, the agents of techno— logical change seek to create new needs by first creating technologies and then trying to create markets for these technologies. This phenomenon is known as a "technology push" situation. Put another way, this is when there is a technological solution in search of a human problem. On the other hand, new technologies may also do a far better job of serving basic human needs which are presently underserved or in fact unserved. In the situation where the problem precedes the solution, this is best described as a "demand pull.‘I Whether a technology is to be accused of serving a demand pull or creating a technology push is somewhat subjective. For some consumer segments a particular technology may be inappropriate, while for other segments it may be highly appropriate. This differentiation task is a basic problem when considering the introduction of an innovation. In a sense, it is a classic marketing problem: does the product come befbre the market or vice-versa? One of the concerns with telecommunication innovations is the extent to which these technologies correspond to the actual needs of the end users. According to the "marketing concept" the needs of a target population would be carefully studied by the entity producing the innovation. The product or service would then be designed and deployed in the market in a manner consistent with the full realiza- tion of a recognized need. This approach is consistent with the demand—pull model of marketing. In the alternative model, technology- push, the needs of the consumer are inadequately considered in the design and marketing of the innovation and may therefore fail to adequately serve these needs. The essential task is to relate measures of human needs to the innovations. Beyond the conceptualization and design stages of tele- communication innovations, there is the problem of trying to predict who will adopt the various telecommunication technologies. In some cases, experience with existing technologies or analogous markets may be useful. It would be useful if the results from studies of telecommunication innovations could be generalized to consider other types of telecommunication innovations. But research findings based on the study of an innovation which has been considered as an indivisible entity may not generalize very well beyond the specific innovation studied. One solution to the problem suggested here is that telecommunica- tion innovations can be considered as divisible units which can be analyzed on the basis of their attributes. The intensity of per- ceived communication needs can be measured and related to expressed preferences fOr attributes of telecommunication technologies. If there is a strong relationship between the two measurement spaces of needs and attributes, one might conclude that a demand-pull model would be most accurate. When a demand-pull exists, the successful diffusion of the innovation might be more confidently predicted. In this case, those consumer segments with the greatest needs would be predicted to adopt those technologies which embody the desired attributes to the greatest extent. Furthermore, it would be expected that these technologies would be applied to serve these same needs. Once a set of needs measures has been related to a set of measures regarding preferred attributes, another relationship can be specified. Initially, the attributes will have been presented in an abstract context divorced from any connection with real world products or administrative entities. Thus, in the next stage, the relationship between the attributes preferred by the consumers and the possibilities for innovations must be considered. If no available technology con- tains the apprOpriate set or amount of the desirable attributes, this would be evidence that underserved or unserved human communication needs existed. A telecommunication innovation could then be developed and positioned to serve these needs. There seem to be two deficiencies in most of the research con- sidering the adoption and diffusion of innovations. First, there has been an undue emphasis on the characteristics of the innovator. Much useful knowledge can be gained by also considering dimensions of the innovations themselves. Another deficiency is the limited nature of most innovation studies. Of the substantial research thus completed, most of these studies have considered the innovation as an indivisible concept. This severely limits the generalizability of the research because findings tend to be "all or none" in terms of relevance to other innovations. Review of the literature The adoption and diffusion of innovations is a well studied topic.1 Adoption studies investigate the process the individual engages when 8 making the decision to adopt or not adopt an innovation. Diffusion studies seek to explain the rate of adoption within a social system. Innovations are products, services or concepts which are perceived as "new" by the potential adopter. Attributes of Innovations "Innovations" or "innovative technologies" are constructs which have been operationalized in different ways. One author proposed that innovations must have one or more of these four properties: (1) new- ness from existing products; (2) newness in time; (3) newness in terms 2 In of sales penetration level; (4) consumer newness to the product. its most general sense, an innovation can be an idea, practice or object perceived as new to an individual. This newness need not be objective. In other words, something which has been around for a long time can still be new to the individual in terms of knowledge, atti- tude or decision to adopt an innovation.3 An innovation is usually considered to be "successful" if it is purchased or used by potential adopters. A practical outcome of innovation research is the opportunity to scientifically predict which innovations will ultimately fail, long before they reach the stage at which they are introduced to the marketplace. This is a significant problem as thousands of new products are introduced to the marketplace each year, many of which end up as failures.4 Recently, substantial research efforts have been made to identify key predictors of product successes and failures.5 On the basis of this research, it can be concluded that understanding user needs, external and internal commmunications, product advantages and marketing 6 efforts are all related to product success. One study determined that three key dimensions could discriminate between successes and failures: (1) product (superiority and user advantages); (2) marketing (knowledge and proficiency of activities); (3) technical/production (synergy and proficiency of activities).7 Thio suggested that the ability to predict adoption of an innova- tion is enhanced if the characteristics of both the potential adopter and the target innovation are studied.8 He considered two construc- tions of compatibility between the adopter and innovation. Symbolic compatibility refers to the intangible attributes of an innovation which are subjectively perceived in an idiosynchratic manner. In contrast to symbolic compatibility, functional compatibility refers to innovation attributes which can be determined in a fairly objective manner. In this sense, these attributes are more tangible. Thio cautions that, in this approach, there is an equal importance between (1) the actor's symbolic interpretation (subjective perception) and (2) the observer's definition of functional requirement (objective assessment). An interesting perspective on the study of innovations in tele- communication is that the adoption process usually requires the adoption of an intangible innovation (e.g. the way one communicates) in addition to a tangible innovation (e.g. the actual product or service). Since the process of communication is central to our functioning, the interaction of telecommunication innovations with the way we communicate is a compelling research consideration. The adoption of a tangible innovation usually involves a behavior such as purchasing or making use of a product or service. The adoption of an intangible innovation on the other hand, requires a symbolic decision. One might accept the cost and convenience of electronic home shopping 10 but reject the implied loss of social interaction due to the inherent nature of the technology. This would then lead to a rejection of the home shopping system, even though its tangible attributes perfectly suited the individual's cost and convenience needs. It is this linking of intangible with tangible innovations that seems to distinguish the study of telecommunication innovations from other diffusion research traditions. It is important to consider not only the functional aspects of telecommunications innovations, but also the symbolic aspects. The functional attributes of innovations are relatively fixed (e.g. objective) characteristics. The symbolic characteristics of innovations are variable subject to individual perceptions. Robertson describes innovations in terms of effects upon con- 9 sumption. A continuous innovation has the least disruptive influence on consumption patterns because the innovation is actually an existing product which has been altered in some fairly minor way. A dynamically continuous innovation more often involves the creation of a new product which has a greater effect on consumption patterns. Finally, a discontinuous innovation has the most disruptive influence on con- sumption patterns by creating previously unknown products. This classification scheme can be fairly arbitrary. For example, is the push-button telephone a continuous or a discontinuous product? In some cases, such as computers or satellites, the classification may be more clear (e.g. discontinuous innovations). The point here is that the adoption of any innovation leads to a change of some sort, whether it be in consumption patterns or in terms of the way one communicates. I 11 Calantone and C00per have used a cluster analysis procedure to devise an empirically based categorization scheme of new product types.10 In contrast to Robertson's fairly arbitrary taxonomy of innovations, the cluster routine was able to generate nine product types. Since this analysis was data based, the outcome of nearly 200 new product offerings was plotted. Thus, each product type had not only its unique description but also an estimate of its likelihood of success. These product types included such innovations as "the innovative mousetrap that really wasn't better," and "the better mousetrap with no market- ing." In their analysis and categorization, Calantone and Cooper examined primarily variables related to the product, marketing and production of the product. A major premise of this paper is that changes in the way we communicate, due to the ad0ption of innovation, may have a great impact on our lives. To reiterate what was said earlier, it is this property of telecommunication innovations which contributes to the growing sense of importance to designing telecommunication systems which are sensi- tive to the needs and desires of human participants. Research on the adoption and diffusion of telecommunication innovations can offer a better understanding of how to design or select telecommunication systems which are functionally and symbolically compatible with the needs of the adopters. Sirbu concluded that in the telecommunication marketplace, successful firms will develop innovative new products which are sensi- tive to human factors of communication needs and capabilities. In addition, these successful firms will need to have fast response capabilities to marketplace forces. This limits the usefulness of highly engineered, capital intensive but inflexible systems.]] 12 Compaine has made the point that successful firms will need to envision their customers as information consumers and not book buyers or television viewers.12 His argument is that the communication need resides in the unique utility of the content and not the conduit through which the information arrives. Compaine argues that new tele- communication technologies are doing little to change the actual content of information, but are making major changes in how easily information can be accessed and processed. Consumers are likely to make major investments in purchasing telecommunication products or services, only to the extent that they gain access to (or can process differently) information in a manner superior to their present capa- bilities. In this light, Frank envisions an interactive database in the home as an example of a defined technology in search of a viable ‘3 Frank indicates that the real challenge is to create application. services that do not duplicate consumer information that the homeowner can get elsewhere, in other forms, at lower costs. Frank finds that even though fuel costs have risen, consumer response has been to plan better and combine trips to save money. The basic desire for mobility has not been eliminated. Finally, Frank perceives that the videotex industry is confident in its ability to develop unique applications that cannot be duplicated easily in other formats. Chaffee and Petrick have suggested several basic parameters salient to a consideration of new technologies.14 They hypothesized that the new technologies will be most socially significant to the degree that they extend our communicatory abilities in terms of (1) storage capacity; (2) access; (3) speed of transmission; (4) amount of information stored or transmitted and (5) reducing information 13 distortion. They argue that since interpersonal communication uses essentially the two senses of sight and sound, technical media innovations which permit the use of other senses of sight and sound, technical media innovations which permit the use of other senses will have only limited social impact. Media which develop the visual illusion of three dimensions (e.g. films of early 1950's) or pursue innovations such as "smell-o-vision" or "feel-o-vision" probably will not make it in the marketplace. They point out that cost is not necessarily a barrier for new technologies, so long as they serve a valid need. Broadcast television receivers were the most expensive consumer oriented technology but also the quickest to diffuse. The attributes suggested by Chaffee and Petrick to describe innovations are fairly technical (e.g. functional). Rogers suggests that it is the attributes of a new product, not as seen by experts, but as symbolically perceived by the potential adopters that really matters.15 In this spirit, he has summarized past thinking and re- search findings to develop a standard classification scheme for describing the perceived attributes of innovations in universal terms. These attributes, while not empirically independent, are conceptually distinguishable: (1) relative advantage; (2) compatibility; (3) complexity; (4) trialability and (5) observability. Relative advantage is the degree to which an innovation is perceived as being better than the idea (product or service) it super- cedes, or which it competes. Compatibility is the degree to which an innovation is perceived as consistent with the existing values, past experiences, present way of doing things and needs of the receivers. Complexity is the degree to which the innovation is perceived as 14 relatively difficult to understand and use. Trialability is the degree to which an innovation may be experimented with one a limited basis. Finally, observability is the degree to which the results of an innovation are visible to others. Theoretically, innovations which are high in all attributes but complexity are most readily adopted. Ostlund found that innovative behavior (e.g. willingness to adopt a product if available) was positively correlated with relative advan- tage, compatibility, trialability, and observability, but negatively with complexity and perceived risk.16 17 Similar findings are suggested by Rice and Rogers. Much of the research summarized by Rogers deals with innovations unrelated to telecommunication. However, some research has been done on telecommunication innovations. Graham for example, suggested that broadcast television diffused more quickly among members of lower social class levels because it was more consistent with the lower- class value system and way of life.18 Dozier and Ledingham used the scheme developed by Rogers and Shoemaker to investigate how people in a cable television market 19 Typical perceive the attributes of interactive cable services. applications of interactive cable television include in-home shopping and banking. A small sample was selected for a focus group interview. Respondents were asked to describe how they currently do things like "keep up with the news, shop, and bank." It was determined who in the household usually handled these tasks. Respondents then had various applications of interactive cable services described to them. These services potentially could do the same type of information tasks (e.g. news, shopping, banking) they were already involved with. Their responses to these deScriptions were recorded and classified. 15 Respondents indicated that interactive cable did have a relative advantage in terms of being able to avoid traffic and waiting lines to conduct informational activities such as banking. But a relative dis- advantage perceived by some of the respondents was a loss of social interaction. Whether interactive cable is perceived as offering relative advantages to current ways of doing things depends on several things. Interactive cable is more efficient than existing alternatives for some uses. If time-saving is an important consideration, inter- active cable is more likely to be perceived as advantageous. The respondents largely perceived interactive cable as a form of "computerization," which had a negative connotation. The theme which emerged was that people did not trust computers and were concerned with "further computer encroachment on individual self-determination and privacy." Attitudes towards computerization seemed to relate to the consequences of possible errors in computer Operations, and loss of a human dimension in conducting transactions such as shopping and banking. Other respondents felt that interactive services would be compatible with their needs and values because of their desire to conduct affairs conveniently. In terms of compatability, a trade-off apparently existed between the negative connotations of "computeriza- tion" and the positive connotations of "convenience." Rice and Rogers made similar findings with respect to computer 20 They discovered that words such as "technol- conferencing systems. ogy," "automation," 'machine," or "computer” tended to increase the resistance of potential adopters to the symbolic idea behind the communication innovation. They suggested that adoption could be in- creased if the convenience of the system was stressed by presenting it as a tool. 16 For the types of service expected to be made available by the cable systems serving the areas in which the respondents lived, the installation and a cost of the service is not a major consideration. Installationconsists of providing a converter and the service would cost about $6.00 per month. Nonetheless, respondents tended to per- ceive the service as fairly complex. They felt that it would require a major modification of their television sets and that it would be extremely expensive to install. Interestingly, several respondents perceived a "generation gap" feeling that they would need to rely on their children to teach them how to use interactive cable services. Children were thus expected to be "early and heavy users." Dozier and Ledingham concluded that, "the dynamics of learning within the house- hold may well prove important in shifting the perception of the innova- tion from one of complexity to one of simplicity. Seeing the informa- tion utility as 'child's play' may well help accelerate the level of system use."21 Experience was also seen to help to resolve inaccurate estimates of the level of complexity involved with operating the system. The final two dimensions considered by Dozier and Ledingham were the trialability and observability of the interactive cable service innovation. They concluded that, "because installation of the con- verter is required to convert a household to two-way cable, and because learning of new skills is required to use the information utility, organized efforts to motivate early adopters to 'show off' the informa- tion utility to neighbors may speed the adoption process." There is a fair amount of empirical support in the diffusion literature to support this notion. Trialability essentially refers to the level and type of risk one is willing to take. The degree of risk perceived relative to 17 the innovation is negatively correlated with innovative behavior.22 0n the one hand there is the risk of not adopting an innovation which has the potential of being very useful and satisfying. On the other hand, there is the risk of adopting an innovation which does not meet expec- tations in some way. For example, the interaction between the functional and symbolic aspects of the innovation nay not appeal to a certain type of person who may seek increased convenience but have a strong need for socializing. Observability seems to appeal to one's sense of vanity. If one does make the commitment to adopt the interactive cable service, there is a tendency to want to "show-off" the results of this decision by inviting friends and neighbors to see the service in action. The appeal of being the "first one on the block" to have interactive cable service is made more attractive if it is possible to show off the service. Innovation Functions In terms of the functional aspects of telecommunication innova- tions, Dozier and Ledingham noted that two distinct types of functions were perceived by respondents. Surveillance functions were described, in which the major activity was information gathering, such that informa- tion was displayed and observed but no modification of a database occurs. Transaction functions involve not only the display but also the modifica- tion of a database, as when bills are paid electronically. Surveillance functions were relatively more attractive than transaction functions to the respondents. The transaction functions seemed to suffer from perceptions of being more complex and less compatible than surveil- lance functions. The point here, is that there seem to be two different types of functional telecommunication innovations which should be con- sidered on their own merits in future research. 18 Sirbu argues that to be successful, an innovation must provide (1) enhanced effectiveness or (2) reduced cost.23 He also posits that it is not just price driven substitution but latent demand expansion that will tend to determine the characteristics of new telecommunication markets. In other words, it will be the new ability to serve the unserved or underserved existing and future communication needs that will play the larger role in the adoption process, while substitution demand based solely on price will be a less important formative force. Because of this, Sirbu suggests that the functional capabilities of each technology be well established and marketed on the basis of these merits. The importance of considering attributes of the innovation is farily well established. These attributes have both symbolic (perceptual) and functional (objective) dimensions. The measurement of the functional attributes can be a fairly straight forward process, as suggested by Sirbu. Investigating the perceptual dimensions is an entirely different problem as noted by Rogers and Shoemaker and others. The five perceptual dimensions suggested by Rogers and Shoemaker are a useful beginning. Two problems are (1) how to measure the symbolic nature of the attributes; and (2) who or what to measure. Perception of Innovations Much of the innovation literature deals with the individual as the unit of analysis. It is also fruitful to consider households, neighbor- hoods or other aggregate units of a social system. Hauser and Koppelman have considered the trade-offs involved in different types of measurement and analysis schemes in dealing with the representation or mapping of perceptual space.24 The three major techniques are (1) factor analysis; (2) similarity scaling; and (3) discriminant analysis. Hauser and Urban also attempted to deal more specifically with the importance of various 19 25 attributes in terms of utility functions. Neslin has attempted with some success to represent relationships existing between product fea- tures (functional) and consumer perceptions of these features.26 Donnelly and Etzel studied variations in the degree of product newness and adoption.27 Their argument was that past research tended to consider genuinely new (e.g. "discontinuous innovation" in Robertson's terminology) as equivalent to superficially differentiated products (e.g. "continuous" or "dynamically continuous" innovations in Robertson's scheme). They also argued that past research tended to be restricted to the consideration of single innovations rather than con- sidering innovations across (or within) product categories, which potentially limits generalizability. Their study investigated the actual purchase of several convenience products of varying degrees of newness to determine whether early adopters of discontinuous innovations were different for the early adopters of continuous or dynamically continuous innovations. Donnelly and Etzel had expert judges rate grocery store products in terms of four objective dimensions. For a product to be considered gen- uinely new, it must be considered new on all four dimensions, other- wise it was considered to be artifically new. One interesting approach used was to consider cognitive styles in terms of how subjects make judgements regarding differences. Pettigrew found that individuals characteristically allow for a certain range of differentiation or "category width" when assessing differences.28 For example, "in selecting maximum values of various optical and auditory phenomena, subjects consistently had either broad, medium or narrow ranges of 29 judgement." Thus, someone who is a "broad categorizer" would tend 20 to judge extreme values in a category (e.g. greater distance from central tendency of a distribution) more often than would someone else who would be described as a "narrow categorizer." The distinction to be made between "broad" and "narrow" categorizers is that if one's cognitive style is to think in terms of broad categories, one tends to assimilate stimuli (differences) in products, essentially overlooking them, recognizing only fairly major changes. Narrow catego- rizers, on the other hand, tend to emphasize any differences. This might mean that continuous or dynamically continuous innovations might appear as "different" and, therefore, "new" to narrow categorizers, but only relatively discontinuous innovations might appear as different and new to broad categorizers. Donnelly and Etzel predicted that individuals with broad category ranges will purchase genuinely new products more frequently than indi- viduals with narrow ranges. They also predicted that individuals with narrower ranges tend to purchase artifically new products more often than those with broader category ranges. They also felt that those in the middle ground would not be more apt to move in either direction in terms of consumption. Although their last hypothesis was actually a test of the null hypothesis, all of their hypotheses were statistically significant. To measure each individual's characteristic category width, a scale developed by Pettigrew was used in this study. Donnelly and Etzel concluded that attributes of the innovation may be as important as behavioral and demographics factors in identifying early adopters. They also found that products could successfully be differentiated in terms of relative newness. 21 Adoption Process Much work has been done which contemplates the adoption process as a process of sequential stages. Mason found that only two stages are necessary and sufficient for adoption to occur, (1) awareness of the innovation and (2) adoption.30 He used a Guttman technique of scalo- gram analysis. His findings indicated that the more complicated sequence of stages in the adoption process which was being preposed by rural sociologists had little merit. He found that adoption processes were not consistent but varied according to the practice studied and the individual farmer, thus no generalizations could be made beyond the two stage model. Several other models have been developed which postulate a series of stages in the adoption process. These stages correspond to a sequence of psychological and behavioral processes which are assumed to be ante— cedants of innovative behavior. There is a fair amount of variation in the number and nature of these stages as identified by different researchers. The traditional view of the innovation decision process recognized by rural sociologists was composed of five states: (1) awareness; (2) interest; (3) evaluation; (4) trial; and (5) adoption. Rogers modified this mode somewhat to describe a five stage process: (1) knowledge; (2) persuasion; (3) decision; (4) implementation; and (5) confirmation.31 The modified model proposed by Rogers resolves some of the deficiencies identified in the traditional rural sociology model. The three major deficiencies were that the five stage model implies that (l) adoption always occurs at some point; (2) the stages occur in order and no stage is skipped; and (3) that once adoption occurs it continues and the innovation is not rejected at a later time. ' 22 Robertson suggests that the exact form of the adoption process will 32 The (vary according to several situational and individual variables. importance of the decision, in terms of consequences of an apparently incorrect decision is a consideration. The extent of meaningful product differentiation and the way the potential adopter categorizes these differences (e.g. narrow vs. broad categories) will have an influence. The extent of the product's conspicuousness (observability) and the consumer's desire for social approval will be related to the decision. Finally, the extent to which the consumer is financially or psycholog- ically able to take risks as well as their decision-making ability will play a part in the adoption process. Rice and Rogers find that, "it seems generally apparent that inno- vations are more successfully adopted when a known or expected demand pulls rather than when the awareness of a new technology pUshes the 33 innovation into the organization." This conclusion that innovations which are matched to intrinsic needs will be more successfully adopted than those which fail to address real needs in a significant way is intuitively appealing. Tauber has found some support for this idea in a nonscientific study using a convenience sample of housewives in prod- uct concept-testing research.34 It seems plausible that adopters who have a relatively high "innate innovativeness" may be more susceptible to a technology push because they are drawn to new ideas as a matter of their basic personality. However, the functional performance of the innovation which they have symbolically accepted may not support a final decision to adopt, due to various situational constraints. One might hypothesize that telecommunication products and services will be most successful if they are useful for serving real needs. The 23 individual's appetite for social interactions may be superceded by a telecommunication system which offers great convenience at a similar or reduced cost. The telecommunication option may not be selected one hundred percent of the time, but to the extent real needs are met, this option nay be more viable. Ideally, the "marketing concept" infers the process of carefully studying consumer needs and desires. Unfortunately, the variable typically studied by market researchers may actually be purchase interest in a product rather than the needs this product can serve.35 The individual's information habits should be studied from the 36 In other words, these "bottom-up" rather than from the 'tOp-down." habits and preferences should be studied from the perspective of the consumer and not the industry supplier. To gain some qualitative in- sights into the way potential adopters may respond to various telecom- munication innovations, Carey reviewed the way people generally approach and use information in their lives by considering the use of telephones and newspapers over time. Carey arrived at several conclusions which may offer some guidance in the planning of telecommunication services. Carey noted in particular, that the development of newspapers and telephone seemed to parallel the way videotex services seem to be emerging, in terms of pricing and human factor barriers. Carey warns that this may signal a slow growth for videotex. Gauging from consumer response to telephone usage and billing, it is probable that people will take advantage of reduced pricing at off-peak hours. Carey indicates that people may prefer flat rate billing to per access charges. 24 Consumers typically may not have much experience in searching through large, complex databases which will be available electronically in the near term. The success of some innovations, such as videotex, may depend on the indexing structure of the software and the extent to which adopters are willing to or can learn how to access specific pieces of information or successfully conduct transactions. Carey finds that consumers have typically been exposed to systems using a simple two step indexing process. In newspapers for example, the first level is a main index to sections followed by a section index to stories. Users of the Prestel videotex service required from six to fourteen indexing 37 An important question to be steps to reach desired information. answered is whether potential adopters will have the patience or the skill to make detailed searches through large databases. Ostlund observed that the emphasis in many studies to date had been forecasting the rate of adoption rather than the prediction of who 38 From a human factors PerSPeCtive’ would or would not adopt and why. it is more interesting to consider the adoption process. Ostlund pointed out that even adoption studies were lacking in their over- emphasis of personal characteristics of adopters. He suggested that attention be paid to perceptual variables relating to the perceived innovation attributes. Ostlund concluded that on the basis of his two studies of several consumer products, the perceptions of innovations by potential adopters are more important predictors of adoption than personal characteristics. Hauser, working specifically with telecom- munication innovations, developed a methodology to study how potential adopters perceive innovations relative to existing technologies and to estimate the relative importance of various attributes of innovations.39 25 Inherent to the "human factors" approach to the development and diffusion of telecommunication technologies is a consideration of the attributes of the innovation. Considering innovations, not as distinct entities, but as collections of characteristics or attributes is not particularly unusual in the innovation literature. However, there has been little done to treat this formally. In the field of economics, Lancaster devised a new approach to consumer theory whereby it was assumed that utility was derived from the attributes of goods and not directly from the goods themselves.40 Lancaster argues that conven- tional economic theory treated goods as single entities and not as collections of characteristics. Lancaster presented the notion that utility or preference orderings are assumed to rank collections of characteristics. In this case, goods are ranked only indirectly, as a function of their attributes. Lancaster described the essence of his new approach to consumer theory by making three points. First, the good itself does not give rise to utility. Second, each good generally possesses more than one characteristic, and many characteristics will be shared by more than one good. Finally, goods in combination may possess characteristics distinct from those pertaining to the goods separately (e.g. synergism). Lancaster's primary contribution was not so much the basic idea of conceiving goods in terms of their attributes, but his advancement of the underlying theory. Hauser estimated the demand and impact of telecommunication inno- vations by positioning these products in perceptual space on the basis of their attributes. In a similar fashion, Quandt and Baumol tried to estimate demand for different modes of transportation by studying the 26 relative importance of different attributes of transportation modes.4] Quandt and Baumol acknowledged Lancaster's work but indicated that their ideas were developed without knowledge of Lancaster's thesis. In essence, Quandt and Baumol hypothesized that it might be use- ful to define a transportation mode in terms of the types of service it provides to the traveler and not in terms of the administrative entity that controls its operations or the sort of physical equipment employed. Modes of transportation could thus be abstractly characterized in terms of several variables salient to transportation, such as: (1) speed; (2) frequency of service; (3) comfort; and (4) cost. By making differ- ential specifications in terms of these four attributes, Quandt and Baumol were able to assess the relative attractiveness of abstract modes of transportation, which nay or may not presently exist. The utility of the model developed by Quandt and Baumol is that the attractiveness of modes of transportation not presently existing can be assessed. Furthermore, the impact of these non-existing trans- portation modes on existing modes can be estimated, should these new modes become available. Quandt and Baumol subjected their hypotheses to a limited test and found encouraging results. The advantages of their model were found to be: (1) the ability to predict for every existing mode of transporta- tion, the effect of introducing a new mode; (2) permitting the hypo- thetical introduction of a new transportation mode simply by specifying its attributes to a sample of potential adopters; and (3) making the forecast of total travel demand a function of the range of travel alternatives. They did note that originally their model assumed ”modal neutrality" meaning that there would be no inherent liking or disliking 27 of particular modes of transportation. In consideration of such factors as "fear of flying" subsequent refinements of the model included binary terms to represent the presence or absence of non-neutral modes. Both Lancaster and Quandt and Baumol have developed formal mathe- matical models to specify the role of innovation attributes in the adoption process. By analogy, their work is clearly applicable to estimating adoption and diffusion of telecommunication innovations. The attributes specified for telecommunication innovations could in- clude the universe of characteristics to be associated not only with existing products, but also potential products with attributes presently unavailable to consumers. In this case, latent or unmet demand can be identified. Collins discussed two types of demand which are influencial in forecasting the use of telecommunication services. The sources of demand are (l) diversion from existing means of telecommunication, e.g. substitution demand; and (2) generation of new traffic, e.g. latent demand.42 Three dimensions can be useful in forecasting the demand for innovative telecommunication services. First, a detailed analysis of present information activities should be undertaken. Next, the suitability of candidate media should be assessed by evaluating their attributes. Finally, the cost effectiveness of each alternative should be weighed. A number of studies have investigated the relationship between adopter characteristics and the rate at which innovations are adopted.43 Three basic categories of variables have been considered: (1) demo- graphics and socio-economic status; (2) personality variables;and (3) communication behavior. Boone for example, found that adopters of 28 cable television service were significantly different from non-adopters on a number of socioeconomic and personality variables.44 Innovators are usually found to be more educated, more venturesome, have favorable attitudes towards credit, are less dogmatic, are better able to deal with abstractions and are more cosmopolitan. In fact, Rogers and Shoemaker have arrived at a set of thirty-two empirically based general- izations about the characteristics associated with adopter categories (e.g. time of adoption).45 Williams and Krugman used Robertson's criteria to establish that public radio was in effect a new product and, therefore, an innovation. They reasoned that since typical descriptors of innovators matched the characteristics of the audience for public radio, then a measure of 45 In fact, this innovativeness should predict public radio listening. was not the case and innovativeness was not a predictor of public radio listening. Other demographic and attitudinal variables were predictors, however. Work which has been done to measure an inherent personality vari- 47 or to locate individuals on a continuum of 48 able of "innovativeness" "adaptiveness" to "innovativeness,' has met with some success. A problem with measuring innovativeness is that this variable seems to be a function of the attributes of the innovation studied and not a com- pletely independent personality variable. Williams and Krugman con- cluded in their study of public radio listening, that their operational- ization of innovativeness may have lacked construct validity. The innovativeness scale they used was developed in product acceptability studies, and they indicated that the link between product purchase decisions and public radio listening is tenuous. 29 Summary In the thousands of innovation studies that have been conducted, scores of variables have been measured and analyzed. Most often the goal of this research has been the successful prediction of innovation adoption. While this is a useful end in itself, there are other possibilities. For example, the diagnostic value of innovation research has not been exploited sufficiently. Merely predicting the rate of adoption (diffusion) or which consumer segments are most likely to adopt an innovation may not be sufficient. In order to maximize the probability of producing an innovation which will be successful, the attributes of the innovations should be deliberately related to the needs of the potential end users. Ideally this would occur in the design stage of the innovation, but may also occur in a later marketing stage. To accomplish this, the kinds of communication needs individuals have and their affective orientation toward the symbolic and functional attributes of innovations should be studied. Telecommunication products and services are likely to be most successful if they are useful for serving existing human communication needs which may be presently underserved or unserved by existing technologies. Two kinds of needs are considered here, those which the user is aware of to some degree, and those needs which the consumer is presently unaware of. If a new technology can serve a set of needs more efficiently than an existing technology, then the older technology will tend to be functionally displaced.49 To the extent that the new technology stimulates the development of "new" needs (e.g. those which the consumer was previously unaware of), the new technology may be more uniquely attractive. It is a difficult research problem to consider 30 measuring needs which the consumer is unaware of. It is useful, how- ever to consider the range of underserved needs. If consumers are not satisfied with what existing telecommunication technologies have to offer them, this knowledge is useful in predicitng who will adopt which technology and why (e.g. to serve which needs). If newer technologies are expected to displace older technologies, these new technologies should be compatible with the needs served by the displaced technology.50 If the newer technologies have tangible attributes which are functionally superior to the displaced technologies (e.g. cheaper, faster), but have intangible attributes (e.g. as sub- jectively perceived by the individual) which are less efficient in serving the needs felt by the potential adopter, these technologies may not survive for long in the marketplace. These human factor con- siderations should be made at all stages of the product life-cycle, 5‘ Inter- from conceptualization to full-scale commercial deployment. estingly, there is actually some concern that such attention to human factors and the marketing concept may actually retard the development of potentially desirable innovations.52 However, this trade-off may well be worth it in the long run. 31 CHAPTER I Reference Notes 1Rogers, Everett M., Diffusion of Innovations, Third Edition, (New York: The Free Press, 1983). Crano, William D., Suellen Ludwig and Gary W. Selnow, Annotated Archives of Diffusion References: Empirical and TheoreticaT Works, prepared by the Center fOr Evaluation and Assessment, MiChigan State University, under the auspices of the U.S. Department of Energy, Office of Utility Systems, (Project No. DE-ACOl-80RG10347), June 1981. 2Robertson, Thomas 5., Innovative Behavior and Communication, (New York: Holt, Rinehart and Winston, Inc., 1971), pp. 4-6. 3Rogers, Diffusion, p. 11. 48002, Allen and Hamilton, Management of New Products, (New York: 8002, Allen and Hamilton, Inc., 1968). 5Marquis, Donald G., "The Anatomy of Successful Innovations," Innova- tion Magazine, November 1969. Davidson, J. Hugh, "Why Most New Consumer Brands Fail," Harvard Business Review, March-April, 1976, pp. 16-24. Kulvik, Hanser, “Factors Underlying the Success or Failure of New Products," Helsinki, Finland, University of Technology. Report No. 29, 1977. 6Calantone, Roger and Robert G. Cooper, "New Product Scenarios: Prospects for Success," Journal of Marketing, Spring 1981, pp.48-60. 7Cooper, Robert 6., "The Dimensions of Industrial New Product Success and Failure," Journal of Marketing, July 1979, pp. 93-103. 8Thio, Alex 0., "A Reconsideration of the Concept of Adapter-Innovation Compatibility in Diffusion Research," Sociological Quarterly, Vol. 12, 1971, pp. 56-68. 9Robertson, 1971, p. 7. 10Calantone and Cooper, 1981. HSirbu, Marvin A., "The Innovative Process in Telecommunications," in Mitchell L. Moss (ed.), Telecommunication and Productivity, (Reading, MA: Adison,Wesley PubliShingiCB}, 198l), pp. 184-198. 32 12Compaine, Benjamin, "Shifting Boundaries in the Information Marketplace," Journal of Communication, Winter 1981, pp. 132-142. 13Frank, Ronald A., "Market Potential of Viewdata Undergoing Tests." Software News, March 1, 1982, pp. 11-14. 14Chaffee, Steven H. and Michael Petrick, Using the Mass Media, (New York: McGraw-Hill, 1975), pp. 231-241. '5Rogers. pp. 211-212. 16Ostlund, Lyman E., "The Role of Product Perceptions in Innovation Behavior," in P.R. MacDonald, Marketing Involvement in Society and the Econom , American Marketing Assoc1ationTProceedings, 1969, pp. 259-266. 17Rice, Ronald and Everett M. Rogers, "Facilitation of Computer Mediated Communication: Innovation in the Organization," International Communication Association, Human Communication Technology Interest Group, Minneapolis, MN, 1981. 18Graham, Saxon, "Cultural Compatibility in the Adoption of Television," Social Forces, Vol. 33, 1954, pp. 166-170. 19Dozier, David M. and John A. Ledingham, "Perceived Attributes of Interactive Cable Services Among Potential Adopters," Inter- national Communication Association, Mass Communication Division, Boston, MA, May 1982. 20Rice and Rogers, 1981. z'Dozier and Ledingham, 1982, p. 17. 22Cunningham, S.M., "Perceived Risk as a Factor in the Diffusion of New Product Information," in Raymond M. Haas (ed.), Proceedings of the American Marketing Association, (Chicago: American Marketing Association,‘l966), pp. 298-721. 23Sirbu, Marvin A., "Innovation Strategies in the Electronic Market- place," Telecommunication Policy, September 1978, pp. 191-210. 24Hauser, John R. and F.S. Koppleman, "Alternate Perceptual Mapping Techniques: Relative Accuracy and Usefulness," Journal of Marketing Research, November 1979, pp. 495-506. 33 25Hauser, J.R. and Glen L. Urban, "Assessment of Attribute Importances and Consumer Utility Functions: Van Neumann-Morgenstern Theory Applied to Consumer Behavior," Journal of Consumer Research, March 1979. pp. 251-262. 26Neslin, Scott, "Linking Product Features to Perception: Self- stated vs. Statistically Related Importance Weights," Journal of Marketing Research, February 1981, pp. 80-86. 27Donnelly, J.H. and M.J. Etzel, "Degrees of Product Newness and Early Trial," Journal of Marketing Research, August 1973, pp. 308-311. 28Pettigrew, Thomas, "The Measurement and Correlates of Category Width as a Cognitive Variable," Journal of Personality, December 1956, pp. 532-544. 29ponne11y and Etzerl, 1973, p. 96. 30Mason, Robert, "An Ordinal Scale for Measuring the Adoption Process," in Studies of Innovators and Communication to the Public, (StanfOrd: ’Institute forTCommunication‘Research, 1962), pp. 101- 116. 31Rogers, Diffusion, p. 164. 32Robertson, pp. 66-75. 33Rice and Rogers, p. 4. 34Tauber, Edward M., "Reduce New Product Failures: Measure Needs as Well as Purchase Interest," Journal of Marketing, July 1973, pp. 61-70. 35Tauber, 1973. 36Carey, John, "Consumer Information Habits," Alternate Media Center, New York University, 1981. 37Carey, 1981, p. 9. 38Ostlund, Lyman E., "Perceived Innovation Attributes as Predictors of Innovativeness," Journal of Consumer Research, September 1974, pp. 23-29. 39Hauser, John R., "Forecasting and Influencing the Adoption of Technological Innovations: Application to Telecommunications Innovations," Working Paper, Transportation Center, Northwestern University, October 1978. 34 40Lancaster, Kelvin, "A New Approach to Consumer Theory," Journal of Political Economy, Vol. 74, 1966, pp. 132-157. 4IQuandt, Richard and William Baumol, "The Demand for Abstract Trans- port Modes: Theory and Measurements," Journal of Regional Science, Vol. 6, 1966, pp. 13-26. 42Collins, Hugh, "Forecasting the Use of Innovative Telecommunication Services," Futures, April 1980, pp. 106-112. 43Rogers and Shoemaker. pp. 185-190. 44Boone, Louis Eugene, The Diffusion of an Innovation: A Socioeconomic and Personality Trait Analysis of Adbpters of CATV Service, Ph.D. dissertation, Uhiversity of AFkansas, 1969. 45Rogers, Everett M. and F. Floyd Shoemaker, Communication of Innovations, (New York: Holt, Rinehart andTWinston, Inc., 1971) pp. l85-l89. 46Williams, Wenmouth and Dean Krugman, "Innovativeness and the Public Radio Audience," Journal of Broadcasting, Winter 1981, pp. 61-69. 47Leavitt, Clark and John Walton, "The Innovative Person and Time of Adoption," Division of Research Working Paper Series, College of Administrative Science, The Ohio State University, 1974. 48Kirton, Michael, "Adopters and Innovators: A Description and Measure," Journal of Applied Psychology, Vol. 61, 1979, pp. 622- 629. 49Owen, Bruce, "The Role of Print in an Electronic Society," in Glen 0. Robinson (ed.), Communications for Tomorrow, (New York: Praeger Publishers, 1978), pp. 229-244. 50Robinson, John P. and Leo W. Jeffries, “The Changing Role of News- papers in the Age of Television," Journalism Monographs, No. 63, September 1979. 5lDucey, Richard V. and Robert E. Yadon, "A Human Factors Research Program for Videotex Technology," IEEE Transactions on Broad- casting, March 1982, pp. 8-19. 52Lawton, Leigh and A. Parasuraman, "The Impact of the Marketing Concept on New Product Planning," Journal of Marketing, Winter 1980, PP. 19-25. CHAPTER II RESEARCH APPROACH Conceptdal Overview This research is an exploratory study of home consumers and the adoption and application of computer based telecommunication technologies. In this research, telecommunication innovations are expected to be per- ceived by potential adopters as additions to an already existing range of options. In some cases, the adopter may use an innovation to func- tionally displace a more familiar way of doing things, because the newer way may be more efficient along one or more dimensions. In other cases, the adopter may acquire new habits or develop new needs when considering the potential applications of an innovation. In any case, options in the telecommunication environment, innovative or otherwise, are selected to the extent that they are able to meet the expectations of individuals. In this study, the potential adopter is seen to respond to an awareness of needs by seeking out stimuli from the environment which are expected to satiate these needs. In the present context, the needs considered are human communication needs and the stimuli are the attributes of various telecommunication options. The conceptual model in Figure 1 illustrates a construction of the adoption process. In this model, the individual develops or becomes cognizant of one or more human communication needs. These needs generate 35 36 a corresponding drive to consider environmental alternatives for satiating the needs. The manner in which these alternatives are ap- proached is derived from the cognitive style of the individual, notably in the fashion by which expectations are structured. Ideally, this model suggests that given a set of needs, the individual is able to generate a sort of mental shopping list of essential attributes which can provide satisfaction. The model depicted in Figure 1 draws upon expectancy value theory, which is well supported in the field of social psychology.1 An applica- tion of this theory has worked well to predict media exposure.2 A rela- tionship was found between gratifications sought and gratifications ob- tained from mass media and exposure to media. In a refinement of mass media uses and gratifications theory, significant empirical support was obtained which differentiated gratifications sought from gratifications obtained.3 This work tends to support the generalization that individuals do seek out specific gratifications or attributes to satisfy needs, and that these attributes may or may not yield obtained gratifications.4 FIGURE 1 CONCEPTUAL MODEL OF THE ADOPTION PROCESS r——— -- — T- ————— 'T -—-- Evaluation 1 . | 1 . ; . t I . s._de___. Determine Select Potential . . , Needs , ideal , technology __, Application ado ter , attributes . 1% 1 e 1 I 1 l 1 l Postpone '—""— '— " ‘L "idecision or not adopt 37 In terms of expectancy value theory, the adoption or non-adoption of computer based telecommunication technologies can be seen to be a function of (l) expectancy and (2) evaluation. Expectancy refers to the perceived probability that an attitude object possesses a particular attribute. Evaluation, positive or negative, occurs after selection and application.5 In an application of expectancy value theory to mass media research to predict program dependency and media exposure, researchers have assumed that media consumers perceive media content in terms of divisible attributes and not ihdivisibly. This approach is maintained in the current study where potential adopters are thought to evaluate telecom- munication alternatives not entirely as indivisible products with labels such as "cable television" or "home computers" but in also terms of their attributes or services. This is consistent with similar work done by Lancaster6 and Quandt andBaumol7 who have found some empirical support for this approach. In the model (Figure 1), once the potential adopter develops a set of desirable attributes (e.g. mental shopping list), the individual seeks to match this ideal set with currently available collections of attributes (e.g. existing products and services). The range of alternatives con- sidered acceptable by the potential adopter is affected by psychological and demographic variables such as knowledge, experience, willingness to take risks, financial status, innovativeness and social contacts. The next step, according to the model, is the selection of a tech- nological product or service which comes closest in attribute space to embodying the most salient attributes. If a selection occurs, a trial application follows. Adoption is not said to occur until after an evaluation, at which point the potential adopter decides whether or not 38 to continue using an innovation, assuming it was an innovation that was selected and not an older alternative. In either case, if the needs are satiated, this information is added to the individual's experiential repertoire. Otherwise, the process may be reengaged. The needs may be redefined in light of experience with what is available, the notion of what the ideal attributes are may change or the actual selection of a technology may change. A similar process is engaged if no initial tech- nology selection was made. The model implies iterative possibilities. The first path through is a "a priori" in the sense that it assumes that an individual develops a set of ideal attributes bgfgrg_actually determining whether these attribures will satisfy the needs. Successive paths through the model are "post hoc" and based on the experience of applying various attributes to serving needs. This transactional model can describe the orientation of a potential adopter to technological additions to the telecommunication environment over a period of time. Of course, it is not very likely that the average person always engages an explicit and conscious decisionmaking process. Selections of technologies may be based on previously formed habits or other constraints. To this extent, the model may be limited. The model does work to deal with these possibilities to some extent by incorporating the possibility of a reduced attribute space being examined in the "ideal attribute" stage which would then lead to a reduced Ntechnologies space." Assuming there is some awareness of a telecommunication innovation, at least two general properties of these innovations may encourage an individual to remove mental cobwebs and become the relatively thoughtful, (11 Dr 39 information-seeking rational consumer considered in the model. First, the innovation is by definition, new to the potential adopter and there- fore remarkable, if only in passing interest. The individual's curiosity may be sufficiently aroused to consider the innovation in greater detail. Another factor is consideration of cost. The cost of computer-based telecommunication innovations can be significant, relative to the house- hold budget. Since a fairly sizeable investment can be made, the potential adopter may be more intent on arriving at a rational decision. This tendency is probably more pronounced as the adoption decision appears more irrevocable. Some attributes of the innovations may initially tempt the potential adopter's imagination and interest leading to a desire to learn more about the innovation. The individual may first learn of an attribute or service only in the context of an innovation. Once aware of this attribute, the individual may go on to develop an appetite or perceived need for this attribute. In a sense, a new need will thus have been created. This form of functional placement is not in conflict with the model, but an exten- sion. It is likely that people will become aware of innovation attributes at some time or other, before they have a sense of a clearly articulated need. This is one of the problems of trying to predict new needs. There is likely to be an interaction between the recognition of salient needs and the awareness of salient attributes. In other words, the potential adopter may not always be able to express a need for something he or she does not know to exist in the external world. However, people may have needs which they can recognize as being presently unserved or underserved. For example, most people would 4O probably like a service that would deliver feature movies to a high definition component color television system in their home on demand, for $1.00 per month. No one offers this service (not surprisingly!), so the "need" is redefined and perhaps a subscription to Home Box Office becomes an acceptable alternative. 0n the other hand, the drive to satiate this communication need may be channeled into something entirely different or perhaps lapse into a latent state. This model can have several potential applications. First, given some knowledge of an individual's needs and their perceptions of various symbolic and functional attributes, a prediction can be made about which _ innovations, if any, are likely to be adopted. Second, some assessment of the impact of the adoption can be estimated. In other words, after adoption, how is the innovation applied? This question can be answered by examining which needs are apparently served. One might operationalize "impact" as the relative extent to which various needs are served. Furthermore, diagnostic insights may be obtained from the use of the model. For example, if a certain set of needs are highly related to salient attributes, one would expect that individual to seek out the innovation which comes closest in attribute space to embodying the pre- ferred set of salient attributes. If no existing innovation can meet this demand, this signals the need for the development of yet another innovation. This type of product positioning may also be useful in testing out product concepts long before full scale committments are made. This sensitivity to human factors may make or break a telecom- munication firm in this increasingly competitive marketplace.8 41 Rationale fOr‘COnducting two stodies As a partial test of the specified model (Figure 1), two studies were conducted. Three measurement spaces were operationalized: (1) salient human communication needs; (2) salient attributes; and (3) appro- priate telecommunication technologies (See Table 1). Statistical rela- tionships were assessed between these measurement spaces. In one study, the relationship between needs and attributes was considered. In the other study, the relationship between these attributes and various tele- communication technologies was explored. An underlying point in the research here, is that by studying human communication needs and the salience of attributes, one can predict which technologies, innovative or otherwise, potential adopters will tend to adopt. The idea of envisioning new products in the abstract as collections of attributes, actually different levels of different attributes, is fairly common in the marketing research literature. This approach to evaluating new product ideas has been used to assess consumer preference for various levels of different attributes for products that do not yet exist. Marketing studies have been successful in asking consumers to make a series of ranking judgements regarding a number of attributes for non-existent products. For example, in considering which attributes a new type of table radio should have, respondents made judgements regarding characteristics such as: alarm type (none, regular, snooze); case-type (pastic, wood); bands received (AM, AM/FM, AM/FM/SW, AM/FM/SW/CB) and cost.9 Generally, this type of marketing research, known as "conjoint analysis," has been popular to test product concepts before the selection of final designs.10 42 .Au ecu m mmuwvcmnam—mu cmzoa 304 mur>cmm :owpsnvgumwu ucwonmupsz mpwppmumm ummucmosa pumgwo mcogawpmh ovum; ammucmogm cowmw>mpop ummuumogm cormv>opmu mpnmu am3103» :o_mw>mpmu mpamu xm31mco pmmgmummgam owcocpumpm gcmpm you've: we’ve 3mw>ugma1>ma mmenmumu pogmcwu payment» mesa; uwcosuumpm aewzccw muse—mm Peas uwcocpompm zgmgnwp owcogpuwpm sewuuagumcw m>wuumgmp=H mcwuo> use: pcwEmmwcmE autos“ memeeee tapes suwpwpa memm omuw> m>euumgwacH agape cmpmczn new mew; mumou cowumucoamcmc» museum mpacmgom chomgma scams >h um~wcmmco mmucmcwm upozmmao: :owumsgoccw ucmccsu mm>wumpmg new mucm'gm saw: xpmh mcvnmmxxooa wavy mam; mpmmxe m>oggeH mew“ Ppwx cmcvmucmucm mm upgoz can .m.: ucmumgwucs cowmcmp weave; use xmpmm mucmggm co we?» mmmb cowumsgomcw ucmscgm>ow :owumsgomcv szmm: mpcm>m ucwgczu ummvmofiocguwh .emppsnwepp< mmummz mmu .H u4mother work which considered present media usage patterns as predictors of interaction with new technologies.8 This previous work differentiated two kinds of communication needs: (1) information-seeking and (2) information habits. Information-seeking is an active process, where people need or want specific pieces of information. Measuring telephone usage patterns seems to be a good indicator and one with some face validity, of the need to conduct information-seeking. Mass media usage tends to be more based on passive habits and less on an active search for information. Therefore, one might predict that the diffusion of computer-based technologies is more likefly to follow the example of the telephone as opposed to technologies like the television for which people seem to have different applications. The major need respondents seemed to be conscious of, with respect to technological alternatives, was the need for managing their schedules. This need was identified as a TIME ELASTICITY factor. The extent to which improvements in convenience are offered by new technologies would seem to be a major consideration in the adoption process. TIME ELASTICITY was a predictor not only of the adoption of innovative but 105 also non-innovative communication products and services. The TIME ELASTICITY factor also predicted the amount of money respondents were willing to pay for a package of computer-based services which were important to them. This study asked respondents to estimate which services they would find important to them and how much they would pay to obtain these services. These operationalizations are somewhat abstract. The validity of data which is based on asking respondents the value of services with which they have no experience is perhaps questionnable. Essentially, the measures in this study were "pre-purchase" measures. The "pre- purchase preference" construct was identified by Danowski and Hanneman in similar research.9 Danowski and Hanneman argued that their data were valid at least to the extent that pre-purchase preferences might predict subsequent purchasing behavior. The present study seems to have supported this supposition. On the basis of the results obtained here, one can predict that TIME ELASTICITY will explain a significant amount of the variance in adoption behavior regarding innovations. One might argue that TIME ELASTICITY measures can predict pre-purchase preferences which in turn can predict actual purchases of innovations. If this line of reasoning is correct, one might also suspect that TIME ELASTICITY should be able to predict adoption of existing products and services. And, in fact, this is the case. Therefore, it appears that more confidence may be placed in the types of attitudinal measures which are used to predict adoption. The fact that respondents had other needs as identified in the factor analyses that did not serve as significant predictors is interesting. The other needs did not predict the number of innovative 106 or non-innovative products adopted nor the amount of money respondents would be willing to pay for computer-based services. Perhaps, respon- dents are limited to considering the new technologies primarily in terms of time elasticity applications. In this case, change agents might do well to consider educating potential adopters to the ways in which new technologies can be applied to serve these other important communica- tion needs.10 CHAPTER IV Reference Notes 1Alpert, Mark I. and Robert A. Peterson, "On the Interpretation of Canonical Analysis," JOUrnal of Marketing Research, February 1971, pp. 67-70. 2Stewart, Douglas and William Love, "A General Canonical Correlation Index," Psychological Bulletin, 1968, v. 70, pp. 160-163. 3Hair, Joseph Jr. et al., Multivariate Data Analysis, (Tulsa, OK: Petroleum Publishing Co., 1979), p. 187. 4Tucker, Raymond K. and Lawrence J. Chase, "Canonical Correlation," in Peter R. Monge and Joseph N. Cappella, Multivariate Techniques in Human Communication Research, (New York: ’Academic Press, 1980), p. 2051228. 5Tucker and Chase, p. 225. 6Cohen, Jacob and Patricia Cohen, Applied Multiple Regression/ Correlation Analysis for the Behavioral Sciences, (Hillsdale, NJ: Lawrence Erlbaum Assoc., 1975), pp. 106-107. 7Dozier, David M. and John A. Ledingham, "Perceived Attributes of Intractive Cable Services Among Potential Adopters," presented to the International Communication Association, Mass Communication Division, Boston, May 1982. 8Carey, John, "Videotex: The Past as Prologue," Journal of Communica- tion, Spring 1982, pp. 80-87. QDanowski, James A. and Gerhard J. Hanneman, "Aging and Preferences for Interactive Cable Services," Journal of Broadcastigg, Summer, 1980, pp. 337-346. 10For a discussion of "Change Agents," see Rogers, Everett M., Diffusion of Innovations, (New York: The Free Press, 1983), pp. 312-346. 107 CHAPTER V METHODS FOR THE TELECOMMUNICATION EXPERTS STUDY Research design The purpose of the second study described in this report was to collect a set of measures to link the attributes and technologies measurement spaces (See Chapter II, Figure 1). A Delphi method was used to survey a group of telecommunication experts who were knowledge- able about technologies used in this field. Members of this group were asked to judge the extent to which an array of telecommunication technologies possessed a number of attributes. Comprehensive instruments (Appendices B and C) were designed to collect data regarding the ability of telecommunication technologies to support various computer-based services. The services described in the Delphi instrument correspond exactly to those in the instrument used in the survey of home consumers. The Delphi method was invented about thirty years ago by re- searchers doing defense work at the Rand Corporation. "Project Delphi" was the name given to a forecasting study sponsored by the United States Air Force. This study employed special techniques to make use of expert opinion in a process of structured group communication.1 Delphi studies typically do not employ random sampling methods. Instead, a universe of "experts" is operationalized and then a purpo- sive sample is drawn from this universe. The sample is selected such that those members of the population who are the most expert with 108 109 respect to some phenomena, tend to be selected. The participants in a Delphi study are asked to make expert judgements regarding some set of phenomena. Individual estimates are then aggregated to produce a group estimate. The goal of the group estimation process in a Delphi study is to use a group of knowledgeable respondents to produce a reliable and valid estimate of an unknown quantity. This quantity might be a physical entity, such as a date; probability of an event; cost; or performance level. The quantity to be estimated might alternatively be an abstract entity such as a normative judgement which identified value structures.2 A rationale for the use of this method is that information which may be more accurate or more objective can be unavailable or prohibi- tively expensive to obtain. The Delphi technique provides an alterna- . tive to forecasting methods whose input parameters are largely subjective individual estimates. The Delphi procedure can be identified in terms of three major components which occur sequentially. First, data are collected from an expert group regarding the estimation of some entity. Second, these data are summarized by the Delphi researcher and the aggregated data are provided as feedback to the expert group. Finally, in light of group feedback, members of the expert group are permitted to reevaluate and perhaps change their original estimates. These basic steps are repeated until there is some kind of consensus or stability (e.g. no change across scuccessive rounds) in this data. Typically, no more than three such iterations will be required.3 The underlying philosophy of the Delphi method is that the judge- ments of individual experts can be improved by exposing each individual to the thoughts of their peers. This process of group communication is 110 structured such that after an initial round of data collection, these data are summarized and then an opportunity is provided for the experts to examine the distribution of estimates. The nature of the feedback can assume various forms. The type of feedback which is presented to group members can affect their responses in subsequent rounds.4 The intent of presenting feedback which summarizes item distribu- tions from the previous round is that once exposed to this information, participants may wish to modify their previous judgements. The feedback process may stimulate any of several reactions by group members. Participants may choose to ignore the feedback and remain with their initial estimates. Group members may also react against the feedback and present a new estimate which is deliberately skewed in an attempt to affect the central tendency of the distribution in a direction they desire. Finally, group members may seek consensus with overall group opinion by revising their original estimates to conform with the central tendency of the distribution. This last outcome is most desirable if the Delphi process is to work effectively. There is some evidence that feedback does in fact tend to stimulate consensus with an expert group.5 The inputs to the Delphi produced group estimates are individual judgements which are derived from a non-probability sample. Researchers commonly apply inferential statistics to data obtained from randomly selected samples. Researchers assume this practice of aggregating data to produce estimates of population parameters is justified. The basis for this justification is an application of probability theory to the sampling process. Since Delphi research does not employ probability sampling some other justification is required for the practice of aggregating data to produce a group estimate. 111 There are several theoretical approaches which have been presented to offer justification and a formal means for aggregating data in Delphi research. The most promising rationale is a "theory of errors" approach.6 Among other things, this approach offers a mathematical rule for deriving a group estimate from a set of individual responses. In the theory of errors approach, individual judgements are treated as though they were a set of readings taken from a single instrument which was subject to random error. In this circumstance, the best estimate of an entity should be a measure of the central tendency of the distribution of obtained readings. Additionally, a measure of dispersion, such as the standard deviation, might be useful to construct a confidence interval about a central value. Data produced from Delphi designs characteristically form lognormal distributions. In this type of distribution, random error is multiplica- tive rather than additive as in Gaussian distributions. In lognomal distributions, the geometric mean is a more accurate estimate of central tendency than the arithmetic mean. For this type of distribution, the median is usually a close substitute for the geometric mean and may be more convenient to use. The theory of errors approach assumes that the judgements of experts is erratic and plagued with random error.7 It also assumes that there is a single underlying "true" parameter which can be estimated by applying human judgement. The theory cannot accomodate the case where there may be two equally valid but different estimates based on different assumptions. In this case, where the distribution may be bimodal, it may be best to proceed as though there were two separate distributions but under mutually exclusive sets of circumstances. If 112 there is no group consensus, this may be due to a totally unreliable set of readings (e.g. expert judgements), or it may indicate that no valid measure can be obtained f0r various reasons. Perhaps the current state of knowledge is insufficient to support any type of consensus. Despite a lack of strong theoretical underpinnings, the theory of errors approach is recommended over other alternatives which have been con- sidered. The theory of errors model can usually provide a better fit of accumulated data to point estimates than these other alternatives. Finally, the theory of errors approach is intuitively attractive because it has the desirable feature of demonstrating the advantage of using the group response over an individual response irrespective of the nature of the physical nature of the process being estimated. In the present research, the theory of errors model is assumed to be operative in the aggregation of individual estimate to produce a group estimate regarding the relationships between attributes and technologies spaces in telecommunication. The number of iterations a Delphi study goes through is a function of the variability of responses and the feedback process. One effect of the Delphi method is a convergence facilitated by the iteration process. Convergence can be defined as the extent to which greater agreement occurs on successive rounds of data collection. One measure of convergence is the changing distance between the upper and lower quartile values for a given item.8 Some would argue that consensus measures do not take advantage of all the information in the distributions.9 According to this line of reasoning, a measure of stability is more informative than consensus measures. When using consensus measures, iterations are continued 113 until a consensus is approximately achieved. This is the operational definition of the best possible group estimate. The use of stability measures would have the iterations continue until the distribution of scores was relatively invariant across two successive rounds. When two successive rounds are similar, even without a con- sensus (e.g. bimodal distribution), this represents the best judgement of the group. As an avenue of further investigation, the reasons for a lack of consensus might be explored. In this study, one iteration was performed. Two major factors contributed to the decision to limit the data collection process to two rounds. First, a visual inspection of the data in Tables XXVII-XXXIV reveals some movement in the central tendencies of the item distribu- tions. The measure of dispersion used, the standard deviation, de- creased in nearly all cases. This presents a convincing case that greater consensus was being achieved in the second round. The second factor is, perhaps, more practical than theoretical. The attrition rate between Rounds I and II was 40%. At this rate, only 6 or 7 respondents would participate in Round II. A decision was made to base findings on the larger group participating in Round 11. According to the theory of errors, the more readings taken, the better, thus it would appear inherently more desirable to use a larger group to produce Delphi estimates. Instrumentation This Delphi survey was conducted by mail. The Round I instru- mentation is presented in Appendix B. The instrumentation for Round II is contained in Appendix C. These appendices include: (1) cover letters; (2) survey instruments; and (3) follow-up letters. 114' The cover letter for the Round I administration identified the nature of the project, its sponsor and the purpose of the research. The process of a typical Delphi study was briefly explained and an example was cited from the research literature with which most respon- dents would be familiar. The output expected from participants was explained. The follow-up letter was sent to everyone in the sample about three weeks after the first mailing to encourage a higher response rate. The survey instruments for the two rounds of data collection were designed to measure three areas. First, each respondent was asked to rate their own technical knowledge of eight telecommunication technol- ogies and different kinds of terminal devices. Second, respondents were asked to estimate the appropriateness of using each telecommunica- tion technology to provide each type of computer-based service. In this section, respondents were also asked to identify the most appropriate attributes for terminal devices. An interval rating scale was used to consider the appropriateness of the match between the services and technologies. A nominal rating scale was used to assess important terminal attributes (see Appendix B for actual items). Finally, each participant was asked to provide some demographic data. Three types of scaling methods for close-ended questions have been popular in Delphi instruments such as those used in the present study: (1) ranking; (2) rating; and (3) paired comparisons. To use the theory of errors model an interval level of measurement is assumed. All three of these scaling methods have been found to exhibit interval level properties. Past Delphi research suggests that respondents tend to prefer the rating method using Likert-type scales. Respondents in 115 this research have indicated that they find rating scales to be fairly comfortable to use in making their estimations. Respondents in Delphi studies have found simple ranking methods to be fairly uncomfortable for them to use. This seems to be because rank- ing requires that no two items can be considered equal on the same dimension. This may be an artificial constraint. In the use of pair comparisons with n items, it can be seen that n*(n-l)/2 comparisons must be made. With just twenty items, this would be 190 separate judgements for participants to make. Obviously, this is time consuming and fatigue effects may influence the quality of the data. As noted previously, in the present study, rating scales were selected for the Delphi instruments. In Delphi research respondents are asked to react to some stimuli presented to them by the researcher. This stimuli might be a situation; a physical event; or some other type of phenomenon. The items in the instrument must present enough detail for the respondent to comprehend the nature of the phenomenon to which she or he is expected to react. It can be a tricky matter trying to present just enough detail to elucidate the problem without providing so much detail that obfuscation becomes inevitable. A lengthy, detailed treatise may tend to obscure rather than reveal the basic premise of the stimuli. On the other hand, too few words may not be sufficient to successfully orient respondents to the phenomena under investigation. Past Delphi research can offer some guidelines. Experience indicates that test items described with lower and higher numbers of words yield the least consensus while items using medium statement lengths produced the highest levels of consensus.10 The problem is trying to provide enough detail so that respondents are able to achieve 116 a common understanding of the stimuli without obscuring the forest for the trees. More words are needed to describe items less familiar to respondents. Statement lengths of twenty to twenty-five words seem to form peak distributions. In the present study, an effort was made to limit statement lengths to these prescribed limits while providing sufficient detail for respondents to make informed estimates. The instruments were pretested,and based on these results, some wording was slightly changed in the final instrument. Selection of the sample As considered in Chapter II, the task for the experts in this Delphi study was to make a set of judgements regarding the extent to which specific telecommunication technologies could support a number of services and features. Thus, the expert group would need to be familiar with the range of both the technologies and the services and features contemplated in the study. The universe of experts operationalized in this study was a special interest group within a major professional association. The Human Communication Technology SpecialInterest Group (HCTSIG) is com- posed of academics and industry professionals who are members of the International Communication Association. The HCTSIG maintains a mailing list of its two hundred-fifty members which is updated annually. This mailing list was used as the sampling frame. As indicated previously, a purposive rather than a probability sampling technique was used in the sample selection process. In this case, the researcher who was a member of the HCTSIG, in consultation with a charter member of the HCTSIG selected from the sampling frame 117 those individuals thought to be most expert in the relevant areas. From a listing of about two hundred-fifty members, forty-nine were initially selected for the sample. The mailing list had been updated after the most recent annual convention of the association which was held in May 1982. At this time, new members and active members had their listings added or verified. An attempt was made to purge inactive listings from the mailing list. There was no prior indica- tion of how successful this effort might have been. Since the function of this Delphi study was to substitute expert judgements for direct knowledge, the relative expertise of the respondents was an important issue. The sample was selected on a basis of limited peer review as well as the fact that individuals self- selected into the special interest group. As an additional check, one section of the instrument was used to collect a set of ratings re- garding the respondents' own self-perceived expertise in various areas. By limiting the final data set to those estimates obtained from respondents who are judged to be expert in the subject matter by not only themselves, but their peers, the accuracy and validity of the results can be substantially improved. The use of self-rating scale has been found to be a useful variable to identify the most expert subsample from within the overall sample initially selected. Accuracy of results was improved by using estimates only from expert subgroups selected in this manner. The accuracy of the results were verifiable by a means external to the study because almanac type questions with known numeric answers were used in that Delphi instrument. In this study, the use of self-rated 11 expert subgroups also enhanced the accuracy of results. In terms 118 of individual responses to items, some further selectivity can lead to superior results. This was found to be true in one study where only those responses which were expressed with a high degree of confidence were included in the summary feedback.12 As might be expected, the greater the number of respondents who felt highly confident, the better the final distribution of group estimations.13 Data regarding the self-perceived expertise of the sample are presented in Table XXV. Respondents were asked to rate their own technical knowledge of nine technologies using a seven point rating scale (1 = not knowledgeable; 7 = knowledgeable). In the responses from Round I, one participant asked that the term "technical knowledge" be further clarified. In the directions for Round II, this construct was operationalized to mean, "the degree to which you are familiar with the capabilities and limitations of each technology described for providing the kinds of services considered in this study." The point was that the respondents were not expected to be capable of designing or building these systems. Respondents, to be considered as experts, were expected to be familiar with technical constraints in terms of applying these technologies to serve human communication needs. In the final analysis of the Delphi data, only those responses from the most expert subgroups were included. Operationally, this meant that data from respondents rating themselves below the midpoint of the expertise scales were not included in the computation of summary statistics. The operationalization of an "expert" to be any respondent rating themselves above the midpoint of the scale is some- what arbitrary. Hopefully, the value of this procedure is evident in terms of trying to identify the most expert subgroup for each technology. 119 Some respondents fit the definition of an expert for some technologies but not for others. In these cases, data were considered from respondents only for those technologies which they had rated themselves as experts. TABLE xxv. DELPHI RESPONDENTS SELF RATINGS 0N KNOWLEDGE OF TELECOMMUNICATION TECHNOLOGIESa FROUND I (Ne20) ROU = Technology Mean (s.d.) Mean (s.d.) One-way cable television 5.15 (1.18) 5.25 (1.82) Two-way cable television 5.00 (1.45) 5.08 (1.73) Broadcast television 5.30 (1.34) 5.00 (2.05) Broadcast radio 5.05 (1.28) 4.97 (1.98) Telephone 5.65 (1.09) 5.58 (1.62) Direct broadcast satellite 4.90 (1.48) 4.92 (1.56) Multipoint distribution service 4.15 (1.90) 3.67 (1.72) Low power television 3.70 (1.84) 3.42 (1.73) Terminals 6.30 (0.92) 5.92 (1.73) a1 = not knowledgeable; 7 = knowledgeable Administration of the Delphi surygy This study was conducted by mailing self-administered question- naires to a sample of forty-nine telecommunication experts identified from a universe of about two hundred-fifty members of the HCTSIG. Two of these questionnaires were returned because the addressee was no longer available at the given listing. Of the remaining forty-seven, 120 twenty-Six people responded to the mailing. Twenty of the returned questionnaires were useable. Three individuals disqualified themselves as experts and three other people felt that the instrument was too long, demanding, confusing or otherwise unworthy of their efforts. The initial mailing contained a cover letter on letterhead stationary and the questionnaire. This mailing went out on August 12, 1982. A follow-up letter urging people to return completed questionnaires was mailed out on September 3, 1982. The initial completion rate of use- able returns to total mailings was 40.8%. Although this was not a random sample, it was desirable to attain the highest completion rate possible to increase the probability that all viewpoints were represented in the data set. There was no way of knowing if those not responding had a systematically different view- point or were in fact disqualifying themselves as experts without communicating this to the researcher. Since the instrument was quite comprehensive and demanding, it was likely that only those with a strong interest were willing to complete the questionnaire. From comments received back from the sample it appears that those who were more certain about the content area appeared to be the ones completing the instrument. Round I data were tabulated and summarized in a convenient form to be presented to the participants as feedback to begin the Round 11 process. Measures of central tendency and dispersion were provided for each item. The Round II questionnaire was sent to participants on October 14, 1982. A follow-up letter was mailed on November 16, 1982. There were twelve questionnaires returned for Round 11. All twelve questionnaires were useable. Since all forty-one members of the 121 initially selected sample who had not disqualified themselves were in- vited to participate in Round II, this yielded an effective response rate of 29.3% for this round. Sample demogrpphics The demographic data for respondents participating in Rounds I and II are presented in Table XXVI. Although the second round of data collection had only twelve participants, compared to the twenty people in Round 1, the demographics for each round are quite similar. The variables measured in this section provide some insight to the type of people who participated in this Delphi survey. The respondents are predominantly from academia holding ranks of assistant, associate or full professor. There was some representation from outside the academic environment. About a third of the partici- pants had received undergraduate training in a communication related area such as radio/television, mass communication or speech. Those with work outside the communication area came from such disciplines as electrical engineering, English, agriculture, sociology and business. At the Master's level, there was a greater concentration in the communication area. Sixty percent of Round I respondents had earned a Master's degree in the communication field. A greater percentage of the Round II respondents (75%) hadaMaster's degree in conmunication. Finally, at the doctoral level, just about two-thirds of the partici- pants had earned a Ph.D. in a communication field. Overall, twenty-one different people participated in the two rounds. From this total, 71.4% had earned a Ph.D. in some field and an additional 9.5% of the sample were doctoral candidates. 122 TABLE XXVI. SAMPLE DEMOGRAPHICS FOR TELECOMMUNICATION EXPERTS IN ROUNDS I AND II OF THE DELPHI STUDY Round I Round 11 (N=20) (N-12) Variable % % Profession Academic 80.0 83.3 Degrees with concentration in communication related area B.A./8.S. 35.0 33.3 M.A./M.S. 60.0 75.0 Ph.D. 65.0 66.7 Sex Male 95.0 91.7 Participated in Round I Yes --- 91.7 Length of membership in HCTSIG (months) i (s.d.) 20.69 (11.44) 16.78 (9.64) Number of memberships in technical organizations 7 (s.d.) 1.10 (1.02) 1.33 (.707) Number of memberships in non-technical organizations 3': (s.d.) 1.55 (1.64) 1.57 (.787) Number of technical journals regularly read i (s.d.) 2.53 (2.32) 2.46 (2.52) Age (years) i (s.d.) 36.45 (7.68) 37.75 (8.50) 123 The comparison of percentages between Rounds I and II may be some- what misleading. Due to the relatively small samples in this study, a small absolute change in a category tends to reflect a larger relative change. In Round 1, for example, each person represents 5% of the total. In Round II, each person represents 8.3% of the total. These data are compared in terms of percentages for ease of relative comparison between and within Rounds. The participants were predominantly male in both rounds. This may reflect a general male bias in the technology area which may exist for various reasons.14 Overall, the HCTSIG membership is about 28% female. The originally drawn sample of forty-nine individuals was about 25% female. This proportion corresponds well to the HCTSIG population. While it is not known why people did not respond to the survey, it might be worth noting that two of the three individuals writing to disqualify themselves as experts were female. Perhaps the male concentration in the final sample can be explained by the somewhat greater tendency for females to disqualify themselves as experts, in spite of their interest in this content area. The average age of the respondents was 36.45 years and 37.75 years, respectively, across the two rounds. Participants had been members of the HCTSIG for 1.72 and 1.40 years across the rounds. The HCTSIG had been in existence for about three years at the time of the survey. Respondents were members in an average of 1.10 (Round I) and 1.33 (Round II) technical organizations. On the average, participants belonged to 1.55 (Round I) and 1.57 (Round II) non-technical organiza- tions. Technical organizations included groups such as the Institute of Electrical and Electronics Engineers (IEEE) and the Association for 124 Computing Machinery (ACM). Non-technical organizations were those such as the International Communication Association (ICA) and the Association for Education in Journalism (AEJ). Respondents regularly read an average of 2.53 (Round 1) and 2.46 (Round 11) technically oriented journals. Examples of these journals would be publications such as IEEE Spectrum; Communications of the ACM or BYTE magazine. Thus, academic and non-academic publications were considered. Chapter V Reference Notes 1Dalkey, N. and O. Helmer, "An Experimental Application of the Delphi Method to the Use of Experts," Management Sciences, April 1963, p. 458. 2Dalkey, Norman C., "Toward A Theory of Group Estimation," in H.A. Linstone and M. Turoff (eds.), The Delphi Method: Techniques and A plications, (Reading, MA: Addison-WeSley, AdVanced Book rogram, , p. 236. 3Linstone and Turoff, p. 229. 4Dalkey, N.D., B. Brown and S. Cochran, "The Delphi Method IV: Effect of Percentile Feedback and Feed-in of Relevant Facts," Rand Corporation, RM-6118-PR, 1970. 5Scheibe, M., M. Skutsch and J. Shofer, “Experiments in Delphi Methodology," in Linstone and Turoff, p. 270. 60a1key, 1975. 70a1key, 1975, pp. 236-261. 8Jones, Chester, "A Delphi Evaluation of Agreement Between Organiza- tions," in Linstone and Turoff, pp. 160-167. 9Scheibe et a1. 10Salancik, J.R., W. Wenger and E. Helfer, "The Construction of Delphi Event Statements," Technological Forecasting and Social Change, 1971, v. 3, pp. 65-73. nDalkey, N., B. Brown and S. Brown, "Use of Self-Ratings to Improve Group Estimates," Technological Forecasting, 1970, v. 1, pp. 283- 291. 12Dalkey, N.C., "The Delphi Method: An Experimental Study of Group Opinion," Rand Corporation, RM-5888-PR, 1969. 125 126 13Dalkey et al., 1970. 14Truxal, Carol, "The Woman Engineer," IEEE Spectrum. Apr11 1983. pp. 58-62. CHAPTER VI RESULTS OF TELECOMMUNICATION EXPERTS STUDY The output from this Delphi survey is expected to be of use in planning new applications for telecommunication technology. For example, two-way cable is a relatively new technology which has re- ceived some notice in the mass media.1 The question is whether poten- tial adopters are attracted to the medium itself as a unit or the attributes of the medium. If potential adopters are attracted primarily by the attributes (services and features) then they will tend to adopt those telecommunication technologies which can best provide these salient features. Consumers typically may not be in a position to judge the relative capabilities of different technologies for supporting their needs and desires. This was the primary function of this Delphi survey, to indicate the extent to which various technologies are appropriate choices for providing the attributes or services consumers most need. In an attempt to remain as parsimonious as possible in the intitial development of this model, the existence of "modal neutrality" is a basic assumption. Of course, the assumption of modal neutrality among modes of telecommunication may be invalid. Pay-per-view movies may be technically feasible to offer via broadcast subscription television (STV), cable television and multipoint distribution service (MDS). From a communication needs perspective the consumer should be indif- ferent as to the delivery mode, so long as the service and cost are 127 128 similar. For some reason, MDS reception may be inherently more attrac- tive to the potential adopter. The rooftop microwave reception antenna may appear as something mysterious and "hi-tech" to the adopter's neighbors. This may appeal to the adopter because it may lend some kind of prestige. This communicability or observability of innovations has generally been,found to be a positive predictor of innovation adoption.2 In future research, more intangible attributes of this nature should be included in research designs. In Chapter IV, a major finding was that Ell.°f the services and features of computer-based telecommunication technologies were useful for serving gomg_of the communication needs people have. In this chapter, an attempt is made to identify which technologies are best suited to providing these attributes. The assumption is that, ceteris paribus, people will tend to adopt the smallest set of technologies which most efficiently serve their communication needs. This set of technologies must also be attractive in terms of their intangible attributes. To the extent that one technology cannot serve all com- munication needs, additional technologies will tend to be adopted. These findings will be presented in terms of the technologies. Thus, each technology will be discussed in terms of its appropriate- ness for the provision of various services and features, as judged by the Delphi respondents. In this discussion, a technology is considered as an appropriate medium for a particular service only if the central tendency (median and arithmetic mean) of the group estimate was above the scale midpoint on a seven point scale, this midpoint is 4.0. The analysis of terminal attributes will be presented in a slightly different context as nominal rating scales were used. Terminal 129 attributes rated as appropriate by 50% or more of the sample were considered to be essential for a particular service. Conclusions are based on the data from the second round of data collection. One-way cable television Each telecommunication technology was described to the Delphi respondents in terms of its service area; bandwidth capacity; direc- tionality (one-way vs. two-way); and message-type (voice, video, data). One-way cable television has a somewhat limited service area of perhaps one or two towns unless the systems are joined through either hard 3 The limited service (physical) or soft (simulated) interconnections. area is an artifact of the regulatory model for cable television. Although cable systems built in the early 1980's have bandwidth capacities of up to 450 MHz, many existing cable systems are capable of substantially less bandwidth due to limitations in amplifier tech- 4 nology. Usually, communication systems capable of handling anything more than 3-4 KHz voice grade signals are called broadband circuits.5 Since the wide bandwidth of cable systems can handle voice, video and high speed data communications, this is truly a broadband communication system. Finally, by definition, one-way cable systems are equipped to handle signals moving in only one direction.6 Using the scale midpoint criterion, only two or three services could be supported by one-way cable television systems (Table XXVII). Pay-per-view movies and color signals can be provided and no special user skills such as needing to learn a computer language, are required. The mean for the "user language" item was below the 4.0 criterion (x = 3.55). But the median was at the criterion level (Md = 4.0), the TABLE XXVII. 130 DELPHI RATINGS OF SERVICES AND APPROPRIATE TECHNOLOGIES: ONE-WAY CABLE TELEVISIONa ROUND I ROUND II Service Median Mean 5.0. Median Mean S.D. Fire/burglar alarm 1.13 1.84 1.89 1.00 1.00 0.00 Video games channel 1.45 2.63 2.17 2.38 2.91 1.92 Utility meter reading 1.13 1.90 2.00 1.11 1.64 1.80 Utility load management 1.13 1.74 1.63 1.19 1.64 1.29 Electronic polling 1.09 1.58 1.58 1.05 1.09 0.38 Interactive education 1.19 1.89 1.78 1.29 1.73 1.49 Special interest database 1.36 2.90 2.42 3.00 3.09 1.38 Electronic mail 1.03 1.06 0.24 1.00 ' 1.00 0.00 Balance inquiry 1.13 1.90 2.00 1.05 1.09 0.30 Electronic funds transfer 1.06 1.37 1.38 1.00 1.00 0.00 General interest database 2.75 3.58 2.76 3.88 3.82 1.54 Pay-per-view movies 4.25 4.05 2.55 5.00 4.27 2.24 Medical alarm 1.13 1.95 2.12 ‘ 1.10 1.18 0.60 Electronic Spreadsheet 1.06 1.16 1.79 1.19 1.36 0.67 Color display 6.80 6.06 1.98 6.67 6.50 0.22 Speed 1.10 1.67 1.78 1.63 1.82 0.99 Graphics 1.36 2.53 2.14 2.75 2.91- 1.70 User language 5.50 4.19 2.45 4.00 3.55 2.54 Alphanumeric. keyboard 1.38 2.00 2.17 1.19 1.73 1.56 Memory/storage 1.11 1.21 0.92 1.00 1.00 0.00 Database management 1.12 1.22 0.94 1.00 1.00 0.00 a1 = very inappropriate; 7 = very appropriate technology 13] distribution was negatively skewed and, therefore. the item was con- sidered to "load" on this technology. Only one other item approached the criterion. This was general interest database service. Both the median and mean were quite near the criterion value (Md = 3.88; i = 3.82). The central tendency measures indicated an increase between rounds of data collection. Based on this trend, the nearness to the criterion value and the re- searcher's subjective assessment, this item was also considered to load on this particular technology. Two-way cable television This telecommunication medium also has a limited service area, due the same constraints which were operative f0r the one-way system. Two- way cable systems are broadband, bidirectional systems which can carry voice, video and high speed data signals. Typically, upstream communi- cations are allocated to the 5-35 MHz portion of the cable spectrum.7 This medium is easily judged to be appropriate for the provision of all the services and features considered in this study (Table XXVIII). Broadcast television Both very high frequency (VHF channels 2-13) and ultra high frequency (UHF channels 14-69) were considered.8 Depending upon several variables subject to regulation such as antenna height, power, frequency and other technical factors such as terrain, the service area of this type of technology approximates a metropolitan scale (e.g. radius of fifty miles). The signal is broadband in nature (6 MHz); one-way and capable of carrying voice, video and data communication. There are limitations on the permissible use of broadcast television TABLE XXVIII. 132 DELPHI RATINGS OF SERVICES AND APPROPRIATE TECHNOLOGIES: TWO-WAY CABLE TELEVISIONa ROUND I ROUND II Service Median Mean S.D. Median Mean S.D. Fire/burglar alarm 6.75 6.17 1.58 6.30 6.20 0.92 Video games channel 6.68 6.17 1.38 6.50 6.40 0.70 Utility meter reading 6.90 6.56 1.25 6.79 6.70 0.48 Utility load management 6.75 5.72 2.19 6.17 6.20 0.63 Electronic polling 6.81 6.39 1.46 6.50 6.50 0.53 Interactive education 6.75 6.44 0.92 6.30 6.30 0.68 Special interest database 6.60 6.28 0.96 6.50 6.40 0.70 Electronic mail 6.42 6.18 1.07 6.10 6.10 0.74 Balance inquiry 6.60 6.06 1.51 6.17 6.10 0.88 Electronic funds transfer 6.68 5.83 1.98 6.00 6.00 0.67 General interest database 6.60 6.33 0.91 6.25 6.00 1.25 Pay-per-view movies 6.81 6.39 1.29 6.67 6.60 0.52 Medical alarm 6.81 6.33 1.50 ' 6.50 6.50 0.53 Electronic spreadsheet 6.00 4.77 2.56 5.33 5.70 0.95 Color display‘ 6.77 6.50 0.89 6.33 6.22 0.83 Speed 6.65 6.24 1.03 5.50 6.30 0.82 Graphics 6.60 5.78 1.73 6.10 6.00 0.94 User language 6.70 6.06 1.44 6.00 5.89 0.93 Alphanumeric. keyboard 6.75 , 6.33 1.23 6.50 6.50 0.53 Memory/storage 5.50 4.56 2.57 4.83 4.60 1.08 Database management 6.00 4.82 2.53 4.50 4.30 1.16 a1 = very inappropriate; 7 = very appropriate technology 133 signals from a regulatory standpoint. However, the Federal Communica- tions Commission (FCC) was considering action in the spring of 1983 which would open the use of the vertical blanking interval for teletext applications.9 This medium was judged to be appropriate for three services and features (Table XXIX). A general interest database; color signals and no special user language were appropriate for braodcast television. Presumably, the Delphi respondents were considering teletext applications for this medium when assessing its appropriateness for the provision of a general interest database service.10 Broadcast radio This medium included both the AM (540-1600 KHz) and FM (88-108 MHz) bands. Like television, the actual service areas of individual stations is dependent upon several regulatory and technical variables such as per- mitted operating power, frequency, antenna height (FM) and soil conduc- tivity (AM). Generally speaking, most radio stations are capable of serving a small metropolitan area. Some channels in the AM band are cleared from interference and allowed to operate at high power levels such that they can serve a several state area. Broadcast radio is generally considered to be narrowband and one-way. Voice, music, data and slow-scan television signals are possible using this medium. Group estimates indicate that this medium is not appropriate for any of the services (Table XXX). This is somewhat surprising and an apparent error since the FCC is moving toward lowering of the regulatory barriers against this medium. Radio subcarriers for example, can be used for utility load management and teletext applications which can support at least a general interest database.n 134 TABLE XXIX. DELPHI RATINGS OF SERVICES AND APPROPRIATE TECHNOLOGIES: BROADCAST TELEVISIONa ROUND I ROUND II Service Median Mean S.D. Median Mean S.D. Fire/burglar alarm 1.00 1.00 0.00 1.00 1.00 0.00 Video games channel 1.23 1.79 1.44 1.21 1.40 0.70 Utility meter reading 1.00 1.00 0.00 1.00 1.00 0.00 Utility load management 1.00 1.00 0.00 1.22 1.40 1.27 Electronic polling 1.13 1.39 1.24 1.00 1.00 0.00 Interactive education 1.14 1.83 1.79 1.21 1.70 1.57 Special interest database 1.36 2.58 2.22 3.50 3.30 1.70 Electronic mail 1.00 1.00 0.00 1.00 1.00 0.00 Balance inquiry ' 1.03 1.05 0.23 1.11 1.20 0.63 Electronic funds transfer 1.00 1.00 0.00 1.00 1.00 0.00 General interest database 1.75 3.21 2.62 4.50 4.20 2.20 Pay-per-view movies 1.36 2.84 2.52 2.00 2.40 1.58 Medical alarm ,l.00 1.00 0.00 1.00 1.00 0.00 Electronic spreadsheet 1.18 1.22. 0.94 1.12 1.30 0.68 Color display 6.77 6.00 2.03 6.67 6.40 0.97 Speed 1.06 1.50 1.47 1.21 1.50 0.98 Graphics 1.40 2.50 2.18 1.33 2.10 1.52 User language 6.00 4.25 3.00 4.00 4.00 2.60 Alphanumeric. keyboard 1.25 1.93 2.12 1.33 1.80 1.55 Memory/storage 1.00 1.00 0.00 1.00 1.00 0.00 Database management. 1.00 1.00 0.00 1.00 1.00 0.00 a1 very inappropriate; 7 = very appropriate technology 135 TABLE XXX. DELPHI RATINGS OF SERVICES AND APPROPRIATE TECHNOLOGIES: BROADCAST RA010a ROUND I ROUND II Service Median Mean S.D. Median Mean S.D. Fire/burglar alarm 1.00 1.00 0.00 1.00 1.00 0.00 Video games channel 1.13 1.47 1.38 1.25 1.40 0.84 Utility meter reading 1.00 1.00 0.00 1.00 1.00 0.00 Utility load management 1.00 1.00 0.00 1.00 1.00 0.00 Electronic polling 1.07 1.13 0.50 1.00 1.00 0.00 Interactive education 1.12 1.69 1.70 1.21 1.70 1.25 Special interest database 1.12 1.56 1.26 1.33 1.70 1.06 Electronic mail 1.00 1.00 0.00 1.06 1.10 0.32 Balance inquiry 1.00 1.00 0.00 1.00 1.00 0.00 Electronic funds transfer 1.00 1.00 0.00 1.00 1.00 0.00 General interest database 1.15 2.12 2.23 2.50 2.40 1.58 Pay—per-view movies 1.00; 1.00 0.00 1.06 1.10 0.32 Medical alarm 1.00 1.00 0.00 ' 1.00 1.00 0.00 Electronic spreadsheet 1.07 1.13 0.50 1.06 1.10 0.30 Color display 1.15 1.29 1.07 1.06 1.10 0.32 Speed 1.07 0.50 1.13 1.13 1.40 0.07 Graphics 1.07 1.38 1.26 1.06 1.10 0.32 User language 5.00 4.13 2.95 3.50 3.30. 2.11 Alphanumeric. keyboard 1.50 1.86 2.18 1.06 1.10 0.32 Memory/storage 1.00 1.00 0.00 1.00 1.00 0.00 Database management. 1.00 1.00 0.00 1.00 1.00 0.00 a1 = very inappropriate; 7 = very appropriate technology 136 Telephone This technblogy, the oldest of those considered here, has a virtually unlimited service area. Using twisted pair wire connections with a 3-4 KHz bandwidth capacity (e.g. narrowband), this medium can carry voice, slow-scan video and data communications. It is a fully switched interactive network for communications between two or more nodes. Telephone communications linking three or more points are usually referred to as "conference calls." This medium was seen to be an appropriate choice for all but three services and features (Table XXXI). Pay-per-view movies, color signals and graphics were not judged to be potential attributes of this tech- nology. The narrow bandwidth does prevent a full motion video service which would eliminate the possibility of pay-per-view movies, unless a hybrid technology (e.g. telephone plus STV for example) was used. How- ever, the telephone is a proven technology for the provision of color signals and graphics. The telephone is a fundamental component in the British interactive videotex system known as Prestel.12 This inter- active videotex service routinely provides color graphics and textual material. Direct broadcast satellites (OBS) The service area for this medium depends primarily on the type of transmission beam used. Most proposed 085 services would use spot beam transmission which would correspond roughly to regions the size of time zones in the contiguous United States.13 This medium is broadband (40 MHz channels) and, therefore, capable of carrying voice, video 1 37 TABLE XXXI. DELPHI RATINGS OF SERVICES AND APPROPRIATE TECHNOLOGIES: TELEPHONEa ROUND I ROUND II Service Median Mean 'S.D. Median Mean S.D. Fire/burglar alarm 6.50 5.90 1.45 6.14 6.09 0.83 Video games channel 5.00 4.90 2.00 4.75 4.73 1.74 Utility meter reading 5.75 5.20 1.82 6.33 5.73 1.85 Utility load management 4.50 4.65 2.23 5.75 5.36 1.86 Electronic polling 6.38 5.47 2.04 6.33 6.09 1.04 Interactive education 5.80 4.90 2.21 6.00 4.82 1.72 Special interest database 5.83 4.50 2.65 5.20 5.00 1.55 Electronic mail 6.38 5.47 2.04 6.08 ‘5.73 1.68 Balance inquiry 6.67 6.25 1.21 6.58 6.36 0.92 Electronic funds transfer 6.50 6.30 0.92 6.58 6.36 0.92 General interest database 6.00 4.95 2.33 5.20 4.73 1.68 Pay-per-view movies 1.09 1.84 2.01 1.29 1.91 1.58 Medical alarm 6.88 6.65 0.81 6.71 6.64 0.51 Electronic spreadsheet 5.00 4.16 2.59 5.33 5.27 1.01 Color display* 1.17 2.25 2.32 1.33 1.90 1.45 Speed 6.68 5.94 1.73 6.13 6.00 1.00 Graphics 3.50 3.78 2.65 1.75 3.64 1.75 User language 6.68 5.61 2.17 6.50 6.10 1.10 Alphanumeric. keyboard 6.60 5.28 2.32 6.13 6.00 1.00 Memory/storage 5.00 4.1 2.77 5.80 5.18 1.78 Database-management 5. 00 4.15 2. 72 4. 75 4.73 1.49 6‘1 = very inappropriate; 7 very appropriate technology 138 (including high definition video) and high speed data communications. As planned, this service would be one-way. Group estimates indicated that 085 is appropriate for pay-per-view; color signals and fewer user skills (e.g. no special languages needed). Although not far off the mark, 085 was not seen to be appropriate for the provision of a general database service (Table XXXII). Otherwise, the group evaluated DBS to be essentially the same as one-way cable television. Multipoint distribution service (MOS) This technology is associated with a fairly limited service area of perhaps 15-20 miles in radius. It is capable of transmitting broad- band signals. with this capacity, voice, video and data communications are possible. Generally, it is configured to be a one-way system, but two-way applications are feasible at least for the business market. The sample estimated that this technology was suited to four services and features (Table XXXIII). These attributes were a special interest database; pay-per-view movies; color signals and no user language. The provision of a general database came close but did not meet the criterion value. The trend between Rounds I and II was f0r the general database to be perceived as less appropriate in terms of MDS. Since this trend had a smaller dispersion in Round 11, it appears that the group was reaching consensus on this estimation. Lowjpower television (LPTV) The LPTV service was approved by the FCC in March 1982. At that time, about 6,500 applications had been made which could potentially mean an additional 4,000 new television stations. The service area 139 TABLE XXXII. DELPHI RATINGS OF SERVICES AND APPROPRIATE TECHNOLOGIES: DIRECT BROADCAST SATELLITESa ROUND I ROUND II Service Median Mean S.D. Median Mean S.D. Fire/burglar alarm 1.07 1.12 0.33 1.00 1.00 0.00 Video games channel 1.11 1.29 0.77 1.42 2.18 1.78 Utility meter reading 1.07 1.12 0.33 1.00 1.00 0.00 Utility load management 1.03 1.06 0.24 1.00 1.00 0.00 Electronic polling 1.07 1.31 0.87 1.00 1.00 0.00 Interactive education 1.17 1.94 1.81 1.11 1.55 1.51 Special interest database 1.21 2.29 2.20 2.75 3.09 1.92 Electronic mail 1.23 2.19 1.94 1.19 1.27 0.47 Balance inquiry 1.07 1.41 1.46 1.00 1.00 0.00 Electronic funds transfer 1.19 1.35 1.46 1.00 1.00 0.00 General interest database 1.35 2.65 2.34 3.33 3.46 1.92 Pay-per-view movies 3.00 3.65 2.78 4.25 3.82 1.83 Medical alarm 1.13 1.47 1.38 1.05 1.09 0.30 Electronic spreadsheet 1.00 1.00 0.00 1.00 1.00 0.00 Color display 6.69 5.77 2.20 6.50 6.50 0.53 Speed 1.12 2.06 2.29 1.63 2.18 1.72 Graphics 1.39 2.75 2.44 3.00 2.82 1.66 User language 6.00 4.29 3.00 4.50 4.50 2.17 Aiphanumeric_ keyboard 1.90 2.39 2.63 1.19 1.55 1.04 Memory/storage 1.19 1.35 1.46 1.00 1.00 0.00 Database management. 1.20 1.38 1.50 1.00 1.00 0.00 a1 - very inappropriate; 7 very appropriate technology 140 TABLE XXXIII. DELPHI RATINGS OF SERVICES AND APPROPRIATE TECHNOLOGIES: MULTIPOINT DISTRIBUTION SERVICEa ROUND I ROUND II Service Median Mean S.D. Median Mean S.D. Fire/burglar alarm 2.00 3.23 2.46 1.50 2.83 2.40 Video games channel 4.25 3.85 2.34 3.67 3.29 2.14 Utility meter reading 2.00 3.31 2.56 1.75 2.43 2.23 Utility load management 2.00 3.23 2.49 2.75 2.86 2.19 Electronic polling 1.50 3.33 2.77 1.38 2.57 2.30 Interactive education 5.50 4.17 2.44 3.50 3.50 2.26 Special interest database 4.75 4.00 2.61 4.00 4.00 2.00 Electronic mail 1.50 3.33 2.57 2.67 3.00 2.19 Balance inquiry 1.31 2.62 2.47 1.50 2.50 2.35 Electronic funds transfer 1.22 2.54 2.50 1.50 2.50 2.35 General interest database 5.63 4.00 2.77 3.83 3.83 1.94 Pay-per-view movies 5.75 4.54 2.70 4.50 4.50 2.59 Medical alarm 1.53 3.00 2.61 1.50 2.33 2.42 Electronic spreadsheet 1.27 1.50 1.73 1.25 1.50 0.84 Color display' 6.58 5.64 1.96 6.50 6.33 1.03 Speed 1.36 3.42 3.00 3.50 3.50 2.07 Graphics . 1.50 3.17 2.66 3.50 3.83 2.04 User language 6.17 5.00 2.68 5.00 4.80 2.49 Alphanumeric keyboard 1.42 3.64 3.04 3.00 3.33 2.34 Memory/storage I 1.55 1.92 2. 25 1.50 1.83 0.98 Database management. 1.55 1.92 2.25 1.25 1.50 0.84 a1 = very inappropriate; 7 = very appropriate technology 141 for an LPTV station would be limited to perhaps a radius of twenty miles.14 Technically, LPTV stations are similar to full power tele- vision stations. The major difference is in permissible operating power, hence the name MM television. The group estimates indicate that LPTV is specially suited to only one service, the provision of color signals (Table XXXIV). One might expect that the same attributes ascribed to full power television would be generalized to LPTV. This was not the case. Terminal design, Terminal devices of some sort are>necessary for most if not all of the computer-based services considered in this study. Four basic types of attributes were comtemplated: (1) addressability; (2) micro- processor equipped; (3) keyboard type; and (4) memory. Group estimates regarding terminal design will be discussed interms of these attributes (Table XXXV). Addressability In telecommunications, addressability refers to the ability to use digital codes to specify the destination of signals. An encoded digital address in the signal must match a similar address in the terminal before the message can be successfully recieved. Only those terminals with digital addresses matching the coded representation in the message will be able to properly receive the message. One applica- tion of this is to enable pay-per-view movies. Only those consumer households with specific digital addresses matching those in the transmitted signal could receive the movie. 142 TABLE xxx1v. DELPHI RATINGS OF SERVICES AND APPROPRIATE TECHNOLOGIES: Low POWER TELEVISIONa ROUND I ROUND II Service Median Mean 5 , 0. Median Mean S. D. Fire/burglar alarm 1,05 1.08 0.29 1.00 1.00 0.00 Video games channel 1.17 1.58 1.44 1.67 1.80 1.10 Utility meter reading 1.05 1.08 0.29 1.00 1.00 0.00 Utility load management 1.00 1.00 0.00 1.17 1.25 0.50 Electronic Polling 1.25 1.46 1.51 1.00 1.00 0.00 Interactive education 1.94 2.36 2.34 2.00 2.40 1.67 Special interest database 1.17 1.75 1.55 2.00 2.20 1.30 Electronic mail 1.10 1.18 0.60 1.00 1.00 0.00 Balance inquiry 1.05 1.08 0.29 1.00 1.00 0.00 Electronic funds transfer 1 .00 1 .00 0.00 1.00 1 .00 0.00 General interest database 1.36 2.33 2.02 3.00 2.80 1.30 Pay-per-view MOVIES 1.25 2.58 2.47 2.25 3.00 2.00 Medical alarm . 1.10 1.33 0.89 ' 1.00 1.00 0.00 Electronic spreadsheet 1.25 1.46 1.51 1.13 1.20 0.45 Color display 5.50 4.70 2.67 6.33 6.20 1.10 Speed 1.44 1.73 1.62 1.33 1.80 1.30 Graphics 1.29 2.27 2.10 3.00 2.80 1.30 User language 2.00 3.80 3.01 3.75 3.60 1.82 Alphanumeric keyboard 1.75 2.20 2.53 1.63 2.00 2.24 Memory/storage 1.27 1.50 1.73 1.00 1.00 0.00 Database management 1.27 1.50 1.73 ' 1.00 1.00 0.00 a1 = very inappropriate; 7 = very appropriate technology 143 The group estimated that addressability was a key attribute for terminals used to provide almost all of the services. Addressability was not judged to be important for the processing of color signals. There was a split on the need for user languages. Apparently this lack of consensus indicates that at least two different types of applications were being considered among members of the sample. One type of application would require a special user language, while the other, perhaps more sophisticated application would require special user skills. Microprocessor A microprocessor is a computer consisting of a central processing unit; memory and an input/output interface. This device may be implemented in one or more chips using large scale integration tech- nology. These devices are relatively cheap. The 16 bit Intel 8088 microprocessor chips sell for under $20 each when purchased in quantity. The cheapness and versatility Of these devices make them cost effective for a multitude of consumer applications. The group estimated that a microprocessor equipped terminal was essential for all but fOur services: electronic-polling; pay-per-view movies; medical alarm; and color display. Electronic-polling and pay- per-view services are interactive services which are seemingly easier to implement using terminals with on-board microprocessors. The data are not clear as to why the sample estimates did not reflect this. Keyboardgtype Two types of keyboards were considered. A numeric keyboard, perhaps similar to the twelve key pushbutton telephone with ten digits 144 and two Special characters might be one type. Variations in design are possible, but the dominant characteristic of numeric keyboards is that primarily numeric responses are enabled. Alphanumeric keyboards on the other hand permit the user to compose messages using a full character set of letters, numbers and speical symbols. An alphanumeric keyboard was indicated to be appropriate for four- teen of the twenty-one services. Three services were explicitly identified as needing only a numeric keyboard: electronic-polling; pay-per-view movies and; medical alarm. There was a split regarding the balance inquiry service. This service is commonly implemented in automatic teller machines using only a numeric keyboard. Memory The final type of terminal attribute considered was memory. Memory size was not specified. The task for the sample was just the estimation of whether or not any local memory was appropriate for the various services. Memory equipped terminals were seen to be useful for fifteen services. The group indicated explicitly that memory was not required for five services. Fire/burglar alarms; utility meter reading; electronic-polling; pay-per-view movies and color display were not seen to require local memory capacity in the terminal. Discussion of results Trying to determine the suitability of various telecommunication media for the provision of different services can proceed in a number of different ways. An individual researcher might make his or her own subjective estimates. Or a clearly defined set of objective criteria 145 TABLE XXXV. SELECTION TERMINAL DESIGN ATTRIBUTES BY ROUNDa Addressable Microprocessor Alphanumeric Memory Service Ia IIb I II I II I II Fire/burglar alarm 88.2 90.0 58.8 60.0 11.8 0.0 23.5 20.0 Video games channel 66.7 90.0 88.9 100.0 72.2 70.0 77.8 90.0 Utility meter reading 87.5 100.0 43.8 70.0 6.3 0.0 25.0 30.0 Utility load management 94.1 100.0 58.8 90.0 35.3 10.0 47.1 50.0 Electronic polling 82.4 90.0 41.2 40.0 52.9 20.0 29.4 20.0 Interactive education 83.3 80.0 83.3 70.0 94.4 80.0 83.3 70.0 Special interest database 83.3 100.0 77.8 70.0 94.4 70.0 72.2 70.0 Electronic mail 93.8 100.0 87.5 100.0 100.0 100.0 93.8 90.0 Balance inquiry 88.9 100.0 55.6 50.0 55.6 50.0 50.0 50.0 Electronic funds transfer: 88.2 90.0 64.7 50.0 76.5 60.0 58.8 60.0 General interest database 76.5 100.0 64.7 80.0 88.2 50.0 70.6 70.0 Pay-per-view movies 80.0 100.0 26.7 30.0 20.0 10.0 33.3 20.0 Medical alarm 81.3 90.0 50.0 30.0 ' 37.5 10.0 43.8 30.0 Electronic spreadsheet 68.8 70.0 81.3 100.0 93.8 90.0 87.5 100.0 Color display- 30.0 12.5 30.0 0.0 40.0 0.0 20.0 0.0 Speed 66.7 90.0 73.3 70.0 80.0 80.0 73.3 60.0 Graphics 64.7 66.7 76.5 88.9 76.5 55.6 70.6 66.7 User language ' 71.4 44.4 78.6 77.8 78.6 55.6 85.7‘ 55.6 Alphanumeric. keyboard 66.7 60.0 73.3 80.0 93.3 100.0 73.3 60.0 Memory/storage 75.0 80.0 87.5 100.0 81.3 100.0 93.8 100.0 Database management. 60.0 70.0 86.7 100.0 93.3 100.0 86. 7 100.0 a(NI = 20; N 12). II = Figures represent percentage of sample selecting each attribute for each service. . bRound I cRound II 146 TABLE XXXV. SELECTION TERMINAL DESIGN ATTRIUBES BY ROUND (continued)a Figures represent percentage of sample selecting serivce. , O. 0. Non- No Only Addressable Microprocessor Numeric, No Memory Service I II I II I II I II Fire/burglar alarm 5.9 10.0 5.9 20.0 17.6 10.0 35.3 70. Video games channel 11.1 10.0 5.6 0.0 22.2 10.0 11.1 Utility meter reading 6.3 0.0 25.0 30.0 12.5 10.0 43.8 60. Utility load management 15.0 0.0 11.8 10.0 15.0 0.0 23.5 30. Electronic polling 15.0 10.0 23.5 20.0 58.8 80.0 29.4 50. Interactive education 5.6 10.0 5.6 0.0 5.6 20.0 0.0 20. Special interest database 5.0 0.0 11.1 20.0 11.1 30.0 11.1 10. Electronic mail 0.0 0.0 0.0 0.0 6.3 0.0 0.0 Balance inquiry 0.0 0.0 16.7 30.0 50.0 50.0 22.2 20. Electronic funds transfer 5.9 10.0 11.8 40.0 29.4 40.0 11.8 20. General interest database 11.8 0.0 11.8 10.0 11.8 30.0 5.9 20. Pay-per-view movies 13.3 0.0 40.0 40.0 53.3 60.0 40.0 60. Medical alarm 12.5 0.0 25.0 20.0' .50.0 70.0 31.3 40. Electronic spreadsheet 12.5 10.0 6.3 0.0 12.5 10.0 6.3 0. Color display 60.0 87.5 60.0 100.0 30.0 25.0 60.0 100. Speed 20.0 0.0 13.3 90.0 20.0 10.0 13.3 20. Graphics 17.6 11.1 11.8 11.1 23.5 22.2 11.8 22. User language 21.4 44.4 14.3 22.2 21.4 33.3 7.1 22. Alphanumeric, keyboard 13.3 20.0 6.7 0.0 13.3 10.0 6.7 20. Memory/storage 6.3 0.0 6.3 0.0 25.0 0.0 0.0 0. Database management. 20.0 10.0 0.0 0.0 20.0 0.0 0.0 0 each attribute for each E: c: c: no no :3 c: c: c: c: c: c: c: c: c: c: c: c: c: c: c: 147 specifying precise engineering, economic and other practical constraints might be enumerated. These criteria might then be systematically applied in a laborious process of evaluating each technology. Ulti- mately, this laborious alternative may be unavoidable in a formal cost benefit analysis, especially if certain specific innovation patterns of industrial innovation are followed.15 In the near term, decisions regarding the design and diffusion of technologies must be made. These decisions need to be based on some kind of infOrmational input. Information can range from the totally subjective to the toally objective, as considered above. Using a Delphi procedure to collect information may offer a compromise between these endpoints of the continuum. Assuming group data are better than individual estimates, as an advocate of the theory of errors might argue, the Delphi method can contribute something to this decision- making process. As noted earlier in the chapter, there are some group estimates which do not appear to be quite on the mark. There are several possible explanations for this phenomenon. First, the group simply may be wrong in their estimate. For example, there is no apparent reason to expect that the telephone should not be an appropriate choice for the delivery of color signals and graphics, especially in light of the Prestel experience. Another explanation is that perhaps the group has identified a complex issue with competing but equally valid solutions that is inadequately measured by the instrument. One might seize upon this type of counter-intuitive finding in future research. This is one of the benefits of the group estimation process and the Delphi method in 148 general, discovering things that the researcher as an individual might not have considered. An additional possiblity is that this particular study has pro- duced somewhat invalid data due to faulty instrumentation. Given the informal feedback from the group, this seems to be a fairly likely reason for some of the apparent discrepancies observed in the data. The questionnaire might have been too long, demanding or ambiguous. The description of each item was accomplished as succinctly as possible but perhaps optimal wording had not been achieved. It did appear from informal comments, that the overall task was understood and that respondents found the general nature of the research to be of interest to them. The items may not have been independent of one another. In other words, factors such as item ordering and fatigue» effects may have some bearing on the final data. Each respondent was asked to make approximately two-hundred sometimes quite involved judgements. In future studies, the items might be placed in a random order on each questionnaire to partially overcome this type of effect. Apart from several estimates which may lack validity or corres- pondence with verifiable data, the study was generally successful. These findings should not be taken as conclusive, but as indicative of general directions. It is quite clear that two-way cable systems and the telphone are the most promising technologies for the imple- mentation of most of the services studied. Other technologies are able to provide subsets of services. To the extent that these other technologies offer cheaper alternatives to two—way cable and telephone systems, these may be adopted by consumers. Since each of the 149 technologies other than two-way cable and telephone support only a subset of the services, it would appear likely that consumers would adopt more than one of these technologies to attain the full comple- ment of services most relevant to their own self-perceived communica- tion needs. CHAPTER VI Reference Notes 1See for example, Nicklein, John W., "Wired City, USA," The Atlantic, February 1979, pp. 35-42. 2Rogers, Everett M., Diffusion of Innovations, (New York: Free Press, 1983), p. 232. 3"Cable Interconnects: Making Big Ones Out of Little Ones," Broad- castin , March 1, 1982, pp. 59-62. This constraint is due to the fact tfiat cable systems must seek franchise agreements which usually are awarded by local governments. See: Baldwin, Thomas F. and 0. Stevens McVoy, Cable COmmunications, (Englewood Cliffs, NJ: Prentice-Hall, 1983), pp. 188-236. 4Baldwin and McVoy. 5Graf, Rudolf F., Modern Dictionary of Electronics, (Indianapolis, IN: Howard H. Sams &’Co., Inc., 1977, p. 91). 6Prior to the Supreme Court's ruling in fCC v Midwest Video (440 US689, 1979), the FCC required most cable systems toibe capable of be- coming interactive. For systems to become activated for two-way communication, upstream amplifiers would need to be added. 7Baldwin and McVoy, p. 57. 8The reader will note that the FCC reallocation of a portion of the TV spectrum comprising UHF channels 70-83 from the broadcasting service to the land mobile service in 1970. (FCC Docket No. 18262, FCC 70-519, May 20, 1970.) 9Federal Communications Commission, Proposed Rule, "Proposed Authoriza- tion of Transmission Teletext by TV Stations," BC Docket No. 81-741, Federal Register, December 14, 1981, pp. 60851-60859. 10For a full discussion of teletext applications, See: Tydeman, John et a1., Teletext and Videotex in the U.S., (New York: McGraw-Hill, 1982 . 150 151 1]See for example: "FCC Authorizes Use of AM Carriers for AM Stations." Communications News, July 1982, p. 18; FCC, Notice of Proposed RUlemaking, "USe of the Subsidiary Communications Aurhorization for Utility Load Management," BC Docket No. 81-352, Federal Register, May 21, 1981, pp. 6-15. 12Sigel, Effren (ed.), Videotext: The Coming Revolution in Home/ Office Information Retrieval,’(White Plains, NY: Knowledge Industry PubTicati ons , 1980) . 13Dassler, A. Fred, "What Are They Doing to Direct Broadcast Satellite Services?" Telecommunications, November 1981, pp. 29-32. An argument for using systems with larger service areas is presented in: Harley N. Radin, “More is Not Better: The Case for Three OBS Service Areas," Satellite Communications, August 1982, pp. 32-36. 14Holsendolph, Ernest, "Here Comes Low Power TV," New York Times, April 11, 1982, p. 6F. 15Abernathy, William J. and Sames M. Utterback, "Patterns of Industrial Innovation," Technology Review, June/July, 1980. CHAPTER VII SUMMARY AND DISCUSSION This research seems to offer four contributions to an understand- ing of the adoption of telecommunication innovations and their applica- tion to communication needs. First, some diagnostic methods and tools have been developed. These methods serve to relate communication needs to telecommunication products and services. Second, the usefulness of considering both the tangible and intangible attributes of telecom- munication innovations has received support. Third, there appears to be some correspondence between present attitudes and past adoption behaviors. This correspondence may support an argument that there is some validity in using present attitudes to predict future adoption behaviors. Finally, some useful experience was gained in conducting the Delphi study. In this chapter, these contributions are further considered. In addition, new directions for this type of research are considered. These suggestions are based on the findings and experiences contained in this report. Diagnostic methods and reinvention The limitations of innovation research which focused only on the process culminating in the adoption decision has concerned researchers. Beginning in the 1970's, the subsequent process of implementation or application of adopted innovations began to interest social scientists. 152 153 The most successful innovations were those which performed well in applied settings, in terms of user needs and demands. Innovation providers or producers were, therefore, concerned with their ability to design and package innovations capable of responding to these needs and demands. Adopters were concerned with their ability to apply innovations to their particular set of circumstances. Typically, the effectiveness of innovations in meeting needs was enhanced through some type of adaption or modification either by the adopters or by the innovation producers. It became important to find predictive and descriptive methods for diagnosing the ability of innovations to meet needs in applied settings. In other words, the innovation process was considered to extend beyond the adoption decision to include an imple- mentation phase. The amount of interest in developing diagnostic methods to assess the extent to which innovations were capable of adequately serving user needs in the implementation or application stage stimulated further work in this area. These diagnostic methods, once developed, could be use- ful both in the initial design phase as well as in later stages of the product life cycle.1 It was observed by researchers that innovations are not adopted into static environments. Social scientists realized that adopters were not just passive accepters of innovations. Instead, adopters came to be considered as active in terms of (l) modifying and adapting innovations; or (2) demanding changes made by the innovation provider. The implementation phase is thus dynamic from either or perhaps both the adopter's and provider's perspectives. These two perspectives are modeled in Figure 2.2 This model is an extension of the model first D D -meeeaweep< mmmueczueoaao Pmmnh amoPocsuwh m ..mmE mew Boga Lmnw>oem pmxcmz mcwsgmpmo muzvogq mcwELmuwo \mcmmz . , a, . - - _ _ . mmmuomm onh<>ozzH mmmuoma zomhaoo< mmmucza onhaoo<\onp<>czzH mzp no ammo: .N mmzwnu 155 presented in Figure 1 (Chapter 2). These processes are dynamic in the sense that the adoption and innovation processes are seen to incorpor- porate feedback from the application and evalution stages. Based upon these results, the appropriateness of an innovation for serving a given set of needs are judged. This experience may lead to modifications in the previous inputs to the model. Essentially, Figure 2 depicts an iterative and heuristic process which can loop until both the adopter and the innovation provider are satisfied that needs and hence market opportunities are adequately addressed. Rogers uses the term, "re-invention," to label the process of applying an innovation to serve needs; evaluating this experience and subsequently seeking to modify an adopted innovation to better serve needs.3 In the extreme case, modification may be insufficient and the innovation may be rejected. For example, the consumer who buys a home computer with only a vague notion of what is needed to operate it or what it can do. Eventually, it may turn out that the computer has insufficient processing power to storage to handle tasks expected of it, or the user may find it too difficult to program. In this case, the computer may not be used at all or it may be used in some fashion which differs from its originally intended implementation. There is evidence that as the complexity of innovations increases, so does the tendency for reinvention.4 This tendency increases when the adopter does not have a very rich understanding of the details of the innovation.5 Innovations with attributes suited to serving a wide variety of needs are likely to be reinvented.6 The manner in which the attributes of the innovation are collected or bundled, also affects the reinvention process. Tightly bundled innovations, with highly 156 interdependent components which must be adopted on an "all or none" basis are much more difficult to reinvent or modify than more loosely bundled innovations.7 As a consequence of being tightly bundled, these innovations are less likely to perform satisfactorily in the implementa- tion stage and are more susCeptible to subsequent rejection. The degree to which reinvention occurs seems also to be a function of the types of needs adopters may have. Narrowly defined needs might be associated with less reinvention of adopted innovations. Adopters who define their needs narrowly may seek single application innovations. Adopters who define their needs more broadly may seek general purpose applications. This classification of adopter-type may be important in predicting the success of innovations. ° As argued earlier in this report, telecommunication innovations may be somewhat unique because of their effect on our most characteristic behavior, communication. The reinvention process seems especially inevitable with telecommunication innovations because these technologies can be applied to such a wide set of needs. In addition, telecommunica- tion innovations such as those studied here, can be loosely bundled. This permits the adopter the freedom to select from a range of options in services and features, those which are most suited to their individual needs. In a field study using telephone and two-way cable technologies to provide videotext services to home consumers, it was discovered that most subscribers wanted a smorgasbord of services that will keep them informed; save time and money; and entertain them. These results are consistent with the findings in the present study. Interestingly, this same study also determined that videotex technologies were perceived by 157 adopters as a "comprehensive home information service," and not just an "electronic newspaper." This finding supports a major premise of the present study, namely that innovations such as videotex are perceived in terms of attributes (e.g. various information services) and not as an intact, indivisible product (e.g. "electronic newspaper").8 An intact, indivisible product is by definition, tightly bundled. Innovation research is a rich areain the literature with studies numbering in the thousands. The adoption process is becoming fairly well understood by researchers. The remainder of the innovatiOn process, which includes the application, evaluation and modification stages (e.g. reinvention) are less well understood. The reinvention stage involves the matching of innovation attributes to needs on an applied and experimental level. Based upon their experiences, adopters may seek to modify (reinvent) an innovation. Either they will attempt to modify the innovation themselves or demand changes be made by the innovation provider who will be looking for market opportunities. Alternatively, adopters may simply reject an innovation as hopelessly inappropriate and perhaps seek an entirely different innovation perceived to be better suited to their needs. ' The importance of understanding the needs of potential adopters and how these needs relate to the attributes of innovations seems clear, not only in terms of the original adoption decision, but also in terms of the reinvention process. The ability to diagnose the potential or actual effectiveness of innovations is contingent upon an understanding of how needs and attributes relate. This is the issue to which much of the present research was addressed. 158 This study sought to maintain generalizability by considering broad based communication needs and unbundled innovation attributes. The nature and salience of the needs and attributes were farily well established in the research literature. While it is important to study these established needs, undoubtedly new needs will emerge in the con- text of these sophisticated computer-based telecommunication technologies. Research designs Should be sensitive to these previously unmeasured needs. If one were to conduct a needs assessment and functional analysis of computer-based telecommunication innovations, perhaps two specific constructs might prove useful to this end. The first construct to consider is that of "functional displacement." Assuming that the potential adopter is cognizant of a set of needs and a set of non-innovative alternatives for satisfying these needs, one or more of these alternatives should be selected. These needs may then be: (1) adequately served; (2) underserved; or (3) unserved, given the properties and limitations of the alternative(s) selected. To the extent that present alternatives leave needs underserved or unserved, innova- tions may appear progressively attractive to potential adopters. If innovations can perform functions Similar to the present methods used more efficiently, then these innovations may be adopted and displace the functions of the previous technology. For example, to handle banking in a convenient manner, people may choose an automatic teller machine rather than a live teller for . conducting transactions. This technology has attributes which permit greater flexibility in terms of geographic access and time scheduling. However, even the automatic teller machine technology may be displaced by‘a service which permits banking to be Conducted electronically from the home. 159 The other construct which may be worthwhile to consider is "functional placement." This would relate primarily to unserved needs but may also be appropriate for considering underserved needs. In this case, the potential adopter is seen to be relatively unaware of a need. Actually, a set of needs may not even exist until after the adoption of an innovation. In other words, the need for new functions is somehow placed into the adopter's personal needs agenda as a consequence of adopting an innovation. Functional placement is an outcome of the reinvention process. Once a technology is made available to adopters, the innovation may be applied to serve needs not previously considered. For example, a home computer may be purchased initially forits entertainment applications. After adoption, the symbol processing and storage power of the machine may become evident to the adopter. At this point, the adopter may enter- tain consideration of new uses such as creating and maintaining elec- tronic databases (mailing lists; household inventories or tax records are examples of this). Perhaps the need to keep timely and organized household records was perceived to be adequately served until the user became aware of the computer's potential. Once aware of a superior alternative, an alter- native which offered greater accuracy, capacity and convenience, the adopter might reevaluate their present situation and decide that their need for good records was in fact unserved (or underserved). Whereas "record-keeping" may have meant cardboard boxes stuffed with receipts tucked away in a closet, after witnessing the power of a computer, the placement of a need for a highly organized and accessible database may occur. Naturally, this need corresponds nicely with what the computer 160 has to offer. In fact, this may be another problem. Adopters may prematurely limit the range of potential alternatives considered to serve needs by defining their needs only in the context of immediately available innovations. Future research should consider not only the relatively broad based needs included in this study, but should go on to consider needs which are more narrowly defined. In this sense, "narrowly defined" is meant to indicate operationalizations of "needs" which are appropriate to a consideration of the attributes of computer-based telecommunication technologies. A start in this direction was attempted in this study. Issues of functional displacement or placement should be considered, especially in the reinvention process. Diagnostic methods and tools should be refined to enhance the ability to predict functional displace- ment or placement applications of adoptions. Such methods and tools might include survey and field studies in applied settings using interviewer or respondent administered instrumentation as well as exposing adopters to actual innovation prototypes. Functional displacement can perhaps best be operationalized as the tendency to shift present activities away from one mode and toward another. Banking is a common activity. One might displace the mode of banking from personal visits to the bank, to using automatic teller machines to using a home banking service. In each case, the same need was being served. More importantly, this need existed prior to the adoption of successive innovations. On the other hand, functional placement can be operationalized as the tendency to initiate new activities as the result of innovation adoption. Household record keeping may have consisted of storing things 161 haphazardly in cardboard boxes in a closet. But the introduction of a personal computer to the home environment may stimulate a need to create and maintain a new record keeping structure which capitalizes on the computer's database management capabilities. Future research designs should consider functional displacement and placement constructs. Needs assessments and the relationships between needs and attributes appear to be central to an understanding of how innovations are implemented in the post adoption phase. A diagnosis of processes in this stage can lead to better adoption decisions by enabling better predictions to be made regarding which innovation may be the best choice, given particular needs. This will also permit bet- ter design and marketing decisions to be made by innovation providers. Tangible and intangible attributes In the previous. section, the utility of considering the relationship between communication needs and innovation attributes was discussed. These attributes can be considered further in terms of their tangible or intangible nature. Tangible attributes are those functions of an innovation, such as its services and features, which can be objectively determined. Intangible attributes are more subjectively determined. For example, a home banking service may enable a variety of reliable and accurate transactions. This can be objectively assessed. The extent to which this technology is perceived to be computerized and how potential adopters feel about computerization is a more subjective issue and, thus, an intangible attribute. ' Given the case where there appears to be a nearly perfect corres- pondence between needs and tangible attributes, but no adoption, there 162 would apparently be some kind of alternative explanation. Naturally, a number of situational or demographic variables might explain this. But even holding these variables constant, the issue may not disappear. A home computer may apparently be the perfect solution to word proces- sing needs and yet not be adopted. The explanation could be that the potential adopter has negative attitudes towards computers in general. Perhaps the person feels intimidated, embarrassed or otherwise threatened by computers. Research findings support this possibility. The origins of these attitudes are typically not experientially based. A hands on trial providing direct experience with the technology may reduce some of the negative attitudes towards computers.9 Understanding how people generally feel about things such as com- puters; convenience; saving money and social interaction may enhance the ability to predict adoption of innovations which may affect these aspects of peoples' lives. Strong relationships were found in this study between intangible and tangible attributes of innovations. Intangible attributes, such as those considered in this study should be operationalized in terms fairly specific to the tangible attributes considered to render the best adoption predictions. Attitudes which are more situation specific tend to be better predictors of subsequent behaviors (e.g. adoption) than more generalized attitudinal measures:l0 Intangible attributes are especially important to consider in tele- communication innovations. As cost and convenience become more important, people may begin to choose telecommunication alternatives for things such as banking, shopping, or information retrieval, rather than going to the bank, store or newstand. This may adversely affect other needs which were coincidentally served in the former way of 163 meeting these needs. Going out to the newsstand may not only accom- plish getting a newspaper, it may also be an excuse to get out of the house, meet people or walk a dog. Present attitudes and future adoption As noted in the previous section, research findings indicate that the more Situation Specific the attitudinal measures are, the better the ability for predicting subsequent behaviors. The best way to test the relationship between present attitudes regarding "pre-purchase preferences" or the salience of innovation attributes and actual adoption decisions is to use logitudinal designs capable of empirically linking these domains. An alternative malongitudinal designs which are expensive and time- consuming, is to use cross-sectional desings such as the present study. The validity of cross-sectional findings may be more suspect than find- ings fromlongitudinal research. However, cross-sectional research may be a wiser choice for the beginning stages of a research program. If one is to have confidence in the ability of cross-sectional attitudinal measures to predict futurg_behaviors, then certainly these measures ought to be able to explain a significant amount of variance in similar pagt'behaviors. Some support for this expectation was obtained in the present study. A flaw was noted in the design. The behavioral measure used in this study was household adoption of innovative and non-innovative telecom- munication related products and services. The attitudinal measures were operationalized at the individual level. Thus, a potential source of extraneous variance was not controlled. An alternative explanation 164 for the mixed findings is that the individual interviewed was not necessarily influential in making household purchasing decisions. Regardless of their attitudes, in this case they would have little or no impact on household adoption behaviors. Future research should operationalize the attitudinal and behavioral measures using comparable units of analysis. The selction of the house- hold as the unit of analysis while somehow weighting the importance of individuals in the household decision-making process is probably the most valid and, therefore, promising avenue to pursue. Delphi technique A person who was reasonably well informed on the basic limitations and capabilities of the telecommunication technologies considered in this study could probably do an acceptable job of estimating the relationship between the attributes and technologies spaces considered in Figure 1 (Chapter 11). However, depending upon the nature of the task, groups are found to produce superior results for intellectual 11 tasks although the time to solution may be greater than for individuals.]2 This was one reason for selecting the Delphi technique in the second study reported here. The other reason was that the nature of the task was such that it seemed more efficient to measure subjective judgements rather than to seek more objective measures and avoid human judgement altogether. Generally, the Delphi technique seemed to produce reasonable out- puts. A few anamolies are apparent in the data. Future research might try to reduce the magnitude of the task expected of the respondents and perhaps provide more direction as to exactly what is expected. If a 165 complex and comprehensive instrument is somewhat unavoidable, it may be helpful in terms of reducing respondent fatigue, to divide the partici- pants into subgroups and then subdivide the task for assignment to sub- groups. Integrating major findings of Consumer and Delphi studies The results of the consumer study reveal several things about the process by which computer-based telecommunication technologies are adopted and applied by consumers to serve their needs. First, it is clear that communication needs are related to the salience of various attributes of these technologies (see Tables XIV-XV). In particular, the need to have a degree of flexibility in managing one's time (e.g. time elasticity) is a useful predictor of both the number of innova- tions adopted and the amount one is willing to pay for a package of computer-based services (see Tables XVII-XVIII). It also appears that the attributes of these technologies are use- ful predictors of their economic value to consumers. Several factors are prominent. Consumers are concerned with those attributes which provide information transaction and status-monitoring capabilities among other things (see Tables XXII-XXIII). Present telephone usage patterns were useful to some extent in predicting the economic value of newer computer-based services (see Table XXI). A final major finding of the consumer study was that intangible attributes of technologies affected the perceived salience of various services and features offered (see Table XXIV). From these data one might suspect that an assumption of modal neutrality might not be valid among consumers. In other words, technologies which would appear to 166 threaten social contact or are perceived as being “computerized" may impress consumers more negatively than other technologies. Loss of social contact and computerization were fOund to be negatively related to the salience of technological attributes (see Table XXIV). The point of the Delphi study was to estimate the extent to which a number of different telecommunication technologies could support various computer-based services. The major finding here was that two- way cable television and the telephone were the most appropriate technologies for providing these services (see Tables XXVIII and XXXI). This is very significant Since there was a strong relationship between the communication needs variables as a group and the technology attri- butes variables as a group. A telecommunication technology which can support most or all of the services is likely to be most successful. To the extent that either the telephone or two-way cable television is perceived as less computerized and less of a threat to social inter— action, the chances are better that consumers will successfully adopt these technologies for the provision of computer-based telecommunication services. The other telecommunication technologies are able to support these services to a much lesser extent. Given that two-way cable television or the telephone are the best telecommunication media for providing these services, another question is what type of home terminal to provide. The Delphi study produced a set of estimates which would indicate that most of the services con- sidered in this study would probably require an addressable terminal which is equipped with an on-board microprocessor and local memory and has an alphanumeric keyboard. The addressability is useful for targeting services to specific subscribers. The microprocessor would 167 permit more sophisticated local (e.g. at home) processing and could make the home terminal more user friendly by adjusting to the user's level of competence. Finally, the alphanumeric keyboard permits full messages using words and other symbols so that textual as well as numeric responses and inquiries can be processed. Returning to the model depicted in Figure l, the overall process suggested in this figure is that home consumers develop or become aware of some set of communication needs. Given the awareness of these needs, the consumers seek out attributes or services which will serve these needs. Those technologies which come closest to embodying the preferred set of attributes will do the best job of meeting the demands or needs. Given the present research, the best overall statement that can be made to integrate these two studies is that the set of communication needs which home consumers perceive to be salient are strongly related to a set of computer-based services which can be most appropriately provided via two-way cable television or the telephone. Other telecommunication media choices would be inferior substitutes. LimitatiOns of this research In the consumer study several limitations are apparent. First, this was a cross-sectional study which severely limits the ability to infer causation in spite of the statistical models used to test the hypotheses. Strictly speaking no temporal ordering of the variables can be specified although such an ordering has been suggested in the model used in this research (Figure 1). Another limitation is a validity question. While a goal of this research has been to refine the process of identifying and measuring 168 variables which contribute to variance in the process of adopting and applying technologies to communication needs, some validity problems must be acknowledged. Describing in just a few words the nature of a computer-based service obviously cannot compare to providing the con- sumer with actual hands-on experience. Thus, even though the reli- ability of the measures can be statistically assessed, the more serious concern of validity can only be subjectively estimated. Another limitation is the implicit assumptionin this study that the respondent is influential in determining household adoption patterns. This may not be the case at all. The prevalence of much unexplained variance in the results presented here may be explained by the fact that data collected with the individual as the unit of analysis are used to predict what may be a household decision. This type of "ecological fallacy" can lead to erroneous conclusions.13 It would probably be a better idea to use the household as a unit of analysis when it does not seem likely that the respondents are the decision- makers. An alternative would be to ask to interview a "household decisionmaker." In the Delphi study several other limitations can be observed. First, group judgments were seen to be incorrect in some places. For example, broadcast radio was not seen to be capable of providing services such as utility load management or a general database and yet given recent action by the Federal Communications Commission regarding the use of the Subsidiary Communications Authorization (SCA) service, it would appear that broadcast radio is expected to be capable of providing these services.14 169 Given the time and effort consumed in conducting a Delphi study, it might be more efficient for a fairly knowledgeable researcher to make his or her own estimations regarding the appropriateness of various telecommunication media to support computer-based services. Perhaps the errors detected in the group judgment are more related to the relative expertise of the sample and not a reflection of the Delphi method itself. In eithercase, for purposes of this study it can be seen that the Delphi respondents did identify two-way cable and the Atelephone as the best telecommunication media for providing a variety of computer-based services. To some extent this may have been a fore- gone conclusion and the expense and effort of conducting the Delphi study may not have been justified. The Delphi study did serve the purpose of producing an empirical and reliable set of estimates. This method of estimation may be more defensible than a strictly subjective set of estimates from even a knowledgeable single researcher. Future directions This type of research is useful for making early warning diagnostics 15 The use of attitu- in the beginning stages of a product life cycle. dinal research to predict future behaviors, such as the adoption of innovations is problematical. Still, this research can be useful as an exploratory tool. Future research can benefit from the experiences gained here in several ways. First, the household or a household decisionmaker as the unit of analysis seems appropriate. Second, the use of abstract questionnaire items to measure attitudes towards the salience of various attributes should be improved. 170 Many field studies have been designed to investigate various rela- tionships between consumers and the technologies designed to serve these consumers.16 These studies which are conducted in laboratory or field settings are providing useful information regarding consumer attitudes towards telecommunication technology after they have been able to experience some of the attributes rather than merely having these attributes verbally described in a survey instrument. This effort should be continued and some of the findings suggested in this study should be reassessed in the field. For example, will the findings regarding the relationships between communication needs and the salience of technology attributes be replicated in the field? Will new factors emerge in the needs or attributes variables? After having some experience with the products and services, will the perceived economic value change? What is the nature of these attitudes over time-- will attitudes be Similar at different levels of experience? These and other questions are better suited to laboratory and field research. The initial questions and perhaps some early indication of how the technol- ogies should be constructed can be broached by sample survey research. But to develop more confidence in these findings an attempt to repli- cate these findings in other research settings is important. Conclusion The goal of this research has been to elucidate the inner workings of the adoption and application of computer-based telecommunication technologies. The major thrust evidenced here was to consider the relationship between communication needs and innovation attributes. This seems to present a fruitful manner for a continuing study of 171 innovations. While this research is cross-sectional, it may be useful for pointing the directions for further exploration using longitudinal designs. This research represents a starting point for future research and not a culmination. Perhaps more questions are left unanswered than answered at this point. In any case, it seems that as we create innova- tive ways for serving our communication and information needs, we also create an attendant responsibility to ensure that these innovations are well suited to human as well as technological factors. Innovative tele- communication technologies which are impressive only from an engineering standpoint and not from an applied human perspective may leave us no better off and perhaps worse off than we were before. The ability to diagnose mismatches between technologies and needs before they happen or perhaps as they happen can help us to build better machines which serve to enhance our abilities to communicate and manage information. CHAPTER VII Reference Notes 1Ducey, Richard V. and Robert E. Yadon, "A Human Factors Research Program for Videotex Technology," IEEE Transactions 0n Broadcasting, March 1982, pp. 8-19. 2This model was first presented in, Ducey, Richard V., "User Preferred Features of Broadcasting Innovations: A Needs-Attributes Based Model of the Adoption Process," paper presented to the IEEE Broad- cast Symposium, Washington, D.C., September 1982. 3Rogers, Everett M., Diffusion of Innovations (3rd ed.), (New York: The Free Press, Inc., 83, pp. - 4Larsen, Judith K. and Rekha Agarwala-Rogers, Re-invention in Adoption: A Study of Community Health Centers, (Palo Alto, CA: American Institutes for Research in the Behavioral Sciences), 1977. 5R0 ers et al., The Innovation Process in Public Organizations, ?Ann Arbor, MI: TDept. of J0urna1ism), 1977. 6Rogers, Everett M., "Re-invention During the Innovation Process," in Michael Radnor et al. (eds)., The Diffusion of Innovations: An Assessment, (Evanston, IL: NorthWestern University, Center far the Interdisciplinary Study of Science and Technology), 1978. 7Koontz, Virginia Lansdon, "Determinants of Individuals Levels of Knowledge, Attitudes Towards, and Decisions Regarding a Health Innovation in Maine," Ph.D. thesis, Ann Arbor, Michigan, University of Michigan, 1976; Eric von Hippel and Stan N. Finkelstein, "Analysis of Innovation in Automated Clinical Chemistry Analyzers," Science and Public Policy, Vol. 6, pp. 24-37, 1979. 8"Games Rank First on Video List," Electronic Media, March 31. 1983. p. 8. 9Hiltz, Starr Roxanne and Murray Turoff, The Network Nation, (Reading, MA: Addision-Wesley, Inc., 1978), pp. 433-434. 172 173 10Fishbein, Martin and Lcek Ajzen, Belief Attitude, Intention and Behavior (Reading, MA: Addison-HeSley,7l975). Icek Ajzen and Martin Fishbein, "The Prediction of Behavior from Attitudinal and Normative Viariables," Journal of Experimental PsychOlogy, v01. 6, 1970, pp. 466-487. Richard W. Anderson andTMarkl . Lipsey, "Energy Conservation and Attitudes Toward Technology," Public Opinion Quarterly, vol. 42, 1978, pp. 17-30. 11Davis, J. Group Performance, (Reading, MA: Addision-Wesley, 1969). 12H. Kelley and J. Thibant, "Group Problem Solving," in G. Lindsey and E. Aronson (eds.), The Handbook of Social PsycholOgy (2nd ed) vol. 4, Group Psycholo y and"Phenomena of InteraCtTOn, (Reading, MA: Aaaision-Wesley, 1969), pp. l-IOI. Babbie, Earl, The Practice of Social Research, 3rd edition, (Belmont, CA: Wadsworth PUblishing Co.,71983), p. 80. 13 14The FCC recently deregulated several provisions regarding the SCA, see Broadcasting, April 11, 1983. 15For a discussion of the "product life cycle," see: Sachs, William S. and George Benson, Product Planning and Management, (Tulsa, OK: Penn Well Publishing Co., 1981), pp. 55-82. 16See for example, Tydeman et al., Teletext and Videotex in the U.S. (New York: McGraw-Hill Publishing Co., 1982). APPENDIX A APPENDIX A TELEPHONE SURVEY INSTRUMENT FOR CONS HER STUDY INNOVATION STUDY TEL. NO. ( ) COLS TC 335 SUMMER 1982 DUCEY 355-7563 I.D. No. ( ) 1-3 FINAL STATUS (1) COMPLETE (2) INCOMPLETE (3) REFUSAL 0F INTERVIE” (4) DISCONNECT (5) JUNK (6) OTHER: Call # INTERVIEWER NAME DATE TIME RESULT (BZ[NA/CB)V BZ=BUSY NA=NO ANSWER 1 _____ _____ CB=CALL BACK 2 ____. .____ 3 .___. .___. 4 __ — INTRODUCTION: “Hello, I'm calling from Michigan State University. We're doinga study in the East Lansing area on new kinds of media, such as two-way cable television and home computers and I have a few questions I'd like to ask a male/female over 18 at this number." 1. First, do you live within the city limits of East Lansing? (1) YES (2) NO (TERMINATE) 2. How many working television sets are in your household?. ( ' ) O-9+ 4 (IF ZERO, GO TO Q.4) (IF NOT ZERO, ASK:) 2a. How many of these are color sets? ( ) 0-9+ 5 3. Does your household subscribe to cable television? (1) YES (2) NO - (9) DK/REFUSED 6 (IF NOT GO T0 0.4) (IF YES, ASK;) 3a. Does your household also subscribe to any of the pay movie services such as Home Box Office, The Movie Channel, Cinemax or Escapade? (1) YES (2) N0 (9) DK/REFUSED 7 3b. If YES, ASK: Which services do your household subscribe to? ( ) NUMBER 8 174. 4. 175 Appendix A INNOVATION STUDY SUMMER 1982 PAGE 2 I would like to know how you feel about several different things. tell me whether you strongly agree. agree. feel neutral, disagree or strong disagree with each of the following statements: It is very inconvenient to go out of the house on shopping tri ps . It is very inconvenient to go out of the house on trips to the bank. If I could do things like shopping and banking without leaving home, I would miss seeing and talking to people I normally meet on these trips. I often try new products or services before my friends or neighbors do. Society is becoming too dependent on computers. Computerized records are a threat to my privacy. Now I have some questions about different needs people some- times say they have. Please tell me how important each need is to you by telling me whether it is very important, important, not very important or not important at all. a. b. The The The The The The The The The The The The need need need need need need need need need need need need to to to to to to to to to to to to keep up with current events. obtain useful information for daily life. keep up with the way government does its job. take less time to run errands out of the house. relax and reduce tension. understand what goes on in the 0.5. and world. be entertained. kill time. improve myself. take less time with household bookkeeping. spend time talking with friends and relatives. have news. sports, weather and other types of current information available when it fits my schedule. 5 <: 5 4 U! as a: as 4: a: as a: a: as so 4» VI Please I 3 I no no no no no no no no no no no NVI wwuwuwwwwww no DA - SDA COLS 10 11 12 13 . 14 15 16 17 18 19 20 21 22 23 24 25 26 176 .Appendix A INNOVATION STUDY SUMMER 1982 PAGE 3 5. (cont) ; .— i m. The need to keep records on household finances organized 4 3 2 and up to date. n. The need to have the television programs I like available 4 3 2 when it fits my personal schedule. 0. The need to cut down transportation costs by not going out 4 3 2 of the house as often. 6. Now I have a few questions about different kinds of things you might do. Please tell me whether you do each thing I mention very frequently, frequently, not very frequently or never: I; k E a. How often do you go to a library? 4 3 2 b. How often do you use automatic bank tellers. machines 4 3 2 where you insert a plastic card and enter a password to make a transaction? c. How often do you buy things over the telephone? 4 3 2 d. How often do you ask for prices of things over the phone? 4 3 2 e. How often do you buy things through the mail? 4 3 2 7. Okay, I'd like to ask you a few questions about how much time you spent yesterday using different kinds of media. a. How long did you watch television yesterday? ( ) MINUTES b. How long did you listen to radio yesterday? ( ) MINUTES c. How long did you spend reading newspapers yesterday? ( )MINUTES d. How many personal, non-business telephone calls did you make yesterday? (_ ) EXACT NUMBER e. How many personal letters did you write last month? ( ) NUMBER 8. There are a variety of new computer-based services which could become available to people like yourself. These services make use of either a computer in your home or connect you to a computer by using telephone lines or cable television lines. I would like to know how important these services are to you. For each service I mention, please tell me whether it is very important, important, not very important or not _‘ important at all. g; __ g; a. How important is a service that would connect a computer 4 3 2 to your home to provide automatic fire and burglar alarms? b. How about a service which would let you play different kinds 4 3 2 of video games from your home? COLS E. z I 27 1 28 l 29 z 1 30 1 31 l 32 ° 1 33 l 34 35—37 38-40 41.4} 44.45 46.47 E z 1 48 1 1,9 177 Appendix A INNOVATION STUDY SUMMER 1982 PAGE 4 I (cont) 3: NVI NIAA no N c. How about a service that would connect a computer to your 4 home to automatically read your utility meters? d. How about a service that would connect a computer to your 4 3 2 home to automatically turn down the hot water heater and the heat or air-conditioning in your house when they are not needed, in order to save money? e. How about a service that would let you vote from home in 4 3 2 local elections or opinion polls? f. How important is a service that would let you take school 4 3 2 courses for college credit, if you wish, where you could ask questions or send answers automatically from your home? 9. How about a service that would let you get special interest 4 3 2 information which could be shown as words and pictures on your television screen? This is the kind of information you mightnormally get by going to the library. h. How important is a service that would let you send or read 4 3 2 messages to other people which could be shown on your tele- vision set instead of using paper? Some people call this electronic mail since it is like the regular mail service in some ways. i. How about a service that would let you do things like find 4 3 2 out the balance of your bank accounts? j. How about a service that would let you pay bills automatically 4 3 2 from your house? k. How about a service that would let you get general interest 4 3 2 information which could be shown as words and pictures on your television set? This would be the kind of information you could usually get by going to a newsstand, for example. 1. How about a pay-per-view movie service that would let you pay 4 3 2 only for those movies that you actually watched. rather than paying a flat rate? m. How about a service which would let you automatically signal 4 3 2 a medical emergency in your home? n. How important is a service which would help you do math 4 3 2 problems like keeping track of household finances. doing taxes, or doing homework? How much per month would you be willing to pay for a package of those computer-based services which are most important to you? Would you be willing to pay: (1) Less than $5 per month, (2) $5-10 per month, (3) $11-20 per month, or (4) more than $20 per month? —I 1 COLS 50 51 52 53 51+ 55 57 58 59 6O 61 62 10. 11. 12. 13. 14. 15. 178 Appendix A INNOVATION STUDY SUMMER 1982 COLS PAGE 5 These services we have just talked about can be provided in slightly different ways. To help us understand your preferences, please tell me how important each of these features are. Please tell me whether each feature is very important, important. not very important or not important at all. _‘ E: SE a. How important is it for you to have all of the words and 4 3 2 l 63 pictures be shown in color on your television set? D. How important is it for you to get infOrmation immediately 4 3 2 l 64 after you decide you want it? c. How important are pictures such as graphs or charts to help 4 3 2 1 65 you understand results from math problems? d. How important is it for you to be able to use these services 4 3 2 l 66 and n2; have to learn a computer language like BASIC or FORTRAN? e. How important is it for you to have a typewriter-like key- 4 3 2 l 67 board which would let you type full messages instead of simple yes/no responses? f. How important is it to be able to keep all of your household 4 3 2 1 68 bookkeeping in a computer and ngt have to save all of the paper records? 9. How important is it to have a useful way to index your 4 3 2 1 69 bookkeeping records so you can get the information you want more easily? For the purposes of this study, it is important for us to know what kinds of things you presently have available in your home. How many working radios do you have in your home? ( ) O-9+ 70 How many operating telephones do you have in your home? ( ) O-9+ 71 How many of these telephones have push-buttons instead of rotary dials? ( ) 72 How many of these telephones did you buy from somewhere other than the telephone company? ( ) 0-9+ 73 Do you have a videocassette recorder in your home? (1) YES (2) NO (9) DK/REFUSED 74 IF YES: 15a. How long have you had this videocassette recorder? ( ) MONTHS 75.77 15b. How many pre-recorded Videocassettes do you have? ( ) NUMBER 78-80 16. 17. 18. 19. 179 Appendix A INNOVATION STUDY SUMMER 1982 513;; Page 6 Do you have a videodisc player in your home? (1) YES (2) N0 (9) DK/REFUSED 81 IF YES: 16a. How long have you had this videodisc player? (________) MONTHS 82-84 16b. How many videodiscs do you have? ( ) NUMBER 85-87 00 you have an audio tape recorder in your home? (1) YES (2) NO (9) DK/REFUSED 88 Do you have a video game machine, like Atari, Intellivision or Odyssey? (1) YES (2) N0 (9) DK/REFUSED 89 IF YES: 18a. How long have you had this video game? ( ) MONTHS 90-92 le1 How many game cartridges do you have? ( ) NUMBER 93.95 00 you have a home computer like the Radio Shack TRS-80 or Apple? (1) YES (2) N0 (9) DK/REFUSED 96 IF NO: 19a. Have you ever used a computer at work, school or somewhere else? (1) YES (2) NO (9) OK/REFUSED 97 IF YES: 19b. How long have you had this home computer? ( ) MONTHS 98-100 19c. How much did you pay for this home computer? (_______) DOLLARS 101-104 19d. How much have you spent on buying Software? ( ) DOLLARS 105-108 19e. How many pre-written programs do you have? ( ) NUMBER 109-111 19f. Do you write programs yourself? (1) YES (2) NO (9) DK/REFUSED 112 199. What are the major ways your computer is used in your household? 113-1141 (EXACT RESPONSE) 19h. Who uses the computer most? (1) SELF (2) CHILDREN (3) SPOUSE 115 (4) BROTHERS/SISTERS (5) PARENTS (6) OTHER: 180 Appendix A INNOVATION STUDY ggg; SUMMER 1982 PAGE 7 20. Do you have a stereo system in your home? (1) YES (2) NO (9) DK/REFUSED 116 21. Does your household subscribe to any of the long distance telephone services which Offer cheaper rates thaTregular Bell Telephone service. such as MCI, Sprint or ITT? (cnzcx ALL THAT AppLy) (1) YES (2) NO (9) OK/REFUSED 117 IF YES: 21a. Which service or services do you subscribe to: (1) MCI (2) ITT (3) SPRINT (4) OTHER (9) DK/REFUSED 118 22. Do you have a Citizen's Band or CB radio in your home or car? (1) YES (2) N0 (9) DK/REFUSED 119 IF YES: 22a. How many channels can it receive? ( ) NUMBER 120 22b. How long have you had this CB radio? (_______) MONTHS 121-123 23. Do you have a set of encyclopedias in your home? (1) YES (2) NO (9) DK/REFUSED 124 24. Is there an intercom system in your home? (1) YES (2) NO (9) DK/REFUSED 125 I have just a few final questions ...... 25. Would you please tell me your age? (_______) YEARS 125-127 26. Are you married? (1) YES (2) NO (9) REFUSED 128 27. Are there children under 12 living in your household? (1) YES (2) NO (9) REFUSED 129 IF YES: 27a.. How many children under 12 are living in your household? (____) NUMBER 130-131 28. Counting yourself, how many people live in your household? (_______) NUMBER 132-133 29. How long have you lived in the East Lansing area? (______) YEARS 134-135 181 Appendix A INNOVATION STUDY SUMMER 1982 Egg; PAGE 8 30. How much education have you completed? (1) 8TH OR LESS (5) COLLEGE DEGREE (2) 9TH-12TH (6) GRADUATE WORK (3) HIGH SCHOOL GRADUATE (7) GRADUATE DEGREE 1’6 (4) SOME COLLEGE (9) REFUSED 31. And finally, is your total annual household income more than $15,000? (--) YES (GO TO 0.3la) (1) NO 31a. Is it more than $25,000? (--) YES (GO TO 0.31b) (2) NO 31b. Is it more than $35,000? (--) YES (GO TO 0.31c) (3) NO 31c. And finally, is it more than $45,000? (5) YES (4) NO (9) DK/REFUSED 13? Thank you very much for your help in conducting this survey. RECORD SEX BEFORE MOVING ON TO NEXT QUESTIONNAIRE: 32. SEX: (1) MALE (2) FEMALE 138 APPENDIX B APPENDIX B INSTRUMENTATION FOR DELPHI STUDY - HAVE I MICHIGAN STATE UNIVERSITY COLLEGE OF COMMUNICATION ARTS AND SCIWCES . EAST LANSING ° MICHIGAS ' 48824 PHD. "NW [N THE MA.5S MEDIA August 12, 1982 Dear Fellow HCT Member: I would like to ask for your assistance in conducting a Delphi-type research study on telecommunication technology. As a member of the Human Communication Technology Interest Group of the International Communication Association, you have already indicated an interest and perhaps special knowledge in this area. This research is in connection with my Ph.D. dissertation and is not officially related to any HCT Interest Group function. One recent example of a Delphi study is Joseph N. Pelton's, “The Future of Telecommunication: A Delphi Survey," Winter 1981 issue of the Jgurnal 9f Communication, pp. 177-189. Essentially. the purpose of a Delphi survey is to create a structure for a group communication process.. The end result of this survey is a collection of best estimates regarding some set of phenomena. This result is achieved by asking an expert group to make an initial set of judgements or estimations. These data are tabulated and sent back to the group members. After examining the distribution of group responses, an opportunity is provided to re-estimate initial estimations. The procedure is repeated until some stability has been achieved. Typically, two or three rounds are sufficient. The purpose of this study is to develop estimates on the extent to which various telecommunication technologies can provide a number of teleservices. A system to provide teleservices would include a telecomunication link, a terminal device and a computer system. In this study the concern is with the telecom- munication link and the terminal device. The home television receiver is assumed to be an appropriate display device. The computer systems used to provide these services can be very complicated and better treated as a separate subject. The Structure of the questionnaire suggests two major system components re- quired to provide these computer-based telecomunication services to home consumers: (l) a telecommunication link and (2) some type of terminal device is required in the home. I am asking you to make judgements on what type of terminal is most appropriate for each service. Simple categories of terminals are used. Also, I would like you to rate the extent to which each telecom- munication technology can be used to support each service. Please mark in the appropriate cell, the codes corresponding to the most appropriate terminal device, and the extent to which each telecomunication technology is an appro- priate medium for each service. In responding to these items, feel free to MSUh-WAaio-WWMfit-W 182 183 Appendix 8 consider new uses of the technologies, such as devoting an entire video channel to data communication, or using the available portions of TV and radio signals to transmit data, as in teletext. I have enclosed a self addressed return envelope. Please use this to send the questionnaire back to me. I would appreciate it if you could respond as soon as your schedule permits. Hopefully this will be within one week of the time you receive this. Since at least two and possibly three rounds may be required, the sooner each round is completed, the better. Your responses will remain anonymous. 1 want to sincerely thank you for assisting me with my dissertation research. I am very pleased to have this Opportunity to work with you all. If things go well, perhaps I will have an opportunity to present some of my findings at next year's meeting of the ICA and HCT in Dallas. Thank you again for your interest and cooperation. With best regards. Richard V. Ducey P.S. If you have any queStions, I can be reached at (517) 353-6410 or 355-8372. RVDzaia Enclosure 184 Appendix B HCT Delphi Survey August 1982 Round One DIRECTIONS: This questionnaire is divided into several sections. The first section provides a few key word descriptors of each technology in terms of (l) geographical service area, (2) bandwidth, (3) one-way vs. two-way and (4) type of communication voice, video. data. In this section you are asked to rate your own technical knowledge with respect to the technologies. In the next section you are asked to estimate the extent to which each telecommunication technology can support each service, and which type of terminal is most appropriate for each service. Finally, some demographic questions are asked to assist in the analysis of these data. I. Self Rating on Technologies: For each technology described, please rate your teEhniCal’knowledge on a seven point scale where 7 = very knowledgeable and l=very unknowledgeable about this technology. Not Knowledgeable Knowledgeable A. ONE-WAY CABLE TV 1 2 3 4 5 6 7 (limited service area--one or two towns; broadband; one-way only; voice, video and data) 8. TWO-WAY CABLE TV 1 2 3 4 5 6 7 (same as above, but data communication can be returned from the home,making it two-way) C. BROADCAST TV (UHF or VHF) l 2 3 4 5 6 7 (metro size service area; 6 MHz; one-way; voice. video. data, one-way) D. BROADCAST RADIO (AM or FM) 1 2 3 4 5 6 7 (metro size to interstate service area; narrowband; voice, slow video; data; one-way) E. TELEPHONE l 2 3 4 5 6 7 (virtually unlimited service area; narrowband; voice. slow video. data; fully switched two-way. conference calls make it n-way) F. 085 (direct broadcast satellite) 1 2 3 4 5 6 7 (service areas may correspond to U.S. time zones; voice, video. data; one-way) G. MDS (multipoint distribution service) 1 2 3 4 5 6 7 (service area limited to one city and inneaiate surroundings; voice. video. data; one-way or two-way capable) H. LPTV (low power TV) 1 2 3 4 5 6 7 (service area might extend to several communities; 6 MHz, voice, video, data. one-way) I. TERMINALS l 2 3 4 5 6 7 (adiressability, numerical keypad, alpha- numeric keypad, microprocessor equipped or not) II. .185 Appendix B Rating appropriateness of technologies for services: Two different scales are used in this section. One scale is used for telecommunication technologies (e.g. columns 1-10) and a different scale is used for terminal type (e.g. column 11). Columns I-IO: Please rate each technology on each service using a seven point scale where 7-very appropriate technology and l-very inappropriate technology to support a given service. These judgements should be based primarily on technical criteria, but you may wish to consider other factors such as economic viability. There should be an entry in columns 1-8 and may be entries in 9,10.‘ Column 11: In the last column, please indicate which features or features (multiple choices allowed) the most appropriate terminal should have. Please use the coding scheme below: 1- addressable 5- non-addressable 2- microprocessor equipped 6- non-microprocessor equipped 3- alphanumeric keyboard 7- numeric keyboard (like telephone) (like typewriter) 4- memory 8- no memory Thus. if the combination 1-2-7 is selected, this would describe a terminal which is addressable, has a microprocessor and uses a numeric keypad. , 1:- NOTE: If there is another technology not mentioned. or a hybrid of existing A service that would connect a computer to your home to provide automatic fire and burglar alarms? A service which would let you play different kinds of video games from your home? A service that would connect a computer to your home to automatically read your utility meters? A service that would connect a computer to your home to automatically turn down the hot water heater and the heat or air-conditioning in your house when they are not needed, in order to save money? A service that would let you vote from home in local elections or opinion polls? A service that would let you take school courses for college credit, if you wish. where you could ask questions or send answers automatically from your home? A service that would let you get special interest information which could be shown as words and pictures on your television screen? This is the kind of information you might normally get by going to the library. technologies (e.g. broadcast TV plus telephone) which you feel are appropriate for one or more services. please indicate this on your questionnaire. TV 5 O 21" .22 E F?” 4: 4: F: .L 33 73 a a a. m to to +4 +4 a: u: CL >.>413 3 5 L’3- m a u o .c “a 3 I '0 “a a. - I I w a o I: L 0 g 8 2 r- «n #3 a: c a an n. >. OPGEI— S—l: Terminal type 186 Appendix B >> A f-l— O ”A w- '0->~4 mm '0 w-Q-d) F-I—>¢U U'PQ DDF-L OU>5 UM Quid-l UU“UU «no. mm: VWF- >3>s¢UCU° VG geese y¢= .OP s,- II‘UQQ’ >LL|JE : L ”man-AH” chunk-02.420!— A service that would let you send or read messages to other people which could be shown on your tele- vision set instead of using paper? Some people call this electronic mail since it is like the regular mail service in some ways. A service that would let you do things like find out the balance of your bank accounts? A service that would let you pay bills automat- ically from your house? A service that would let you get general interest information which could be shown as words and pictures on your television set? This would be the kind of information you could usually get by going to a newsstand, for example. A pay-per—view movie service that would let you pay only for those movies that you actually watched. rather than paying a flat rate? A service which would let you automatically signal a medical emergency in your home? A service which would help you do math problems like keeping track of household finances. doing taxes, or doing homework? Capability to have all of the words and pictures be shown in color on your television set? Capability to get infbrmation immediately after you decide you want it? Capability to provide pictures such as graphs or charts to help you understand results from math problems? Capability to use these services and ggt_have to learn a computer language like BASIC or FORTRAN? Capability to have a typewriter-like keyboard which would let you type full messages instead of simple yes/no responses? Capability to keep all of your household book- keeping in a computer? records so you can get the information you want more easily? Provides auseful way to index your bookkeeping ji .13456154lo u III. 187 Appendix 8 Demographics A. B. How long have you been a member of HCT? (months) What is your academic rank or professional title? ( ) Please list the academic areas in which you have earned degrees. gem Arse B.A. H.A./".5. Ph.D. Other In addition to ICA, with what other professional associations do you affiliate? SCA _________AEJ IEEE MAPOR ACM OTHER (please specify) What technically oriented Journals, if any, do you read? (Please list) Age: '(years) Gender: (1) Female (2) Male -4- 188 Appendix B MICHIGAN STATE UNIVERSITY DEPARTMENT OF TELECOMMUNICATION EAST LANSING ' MICHIGAN ' 48824 September 3, 1982 Dear Fellow HCT Member: About two weeks ago you should have received a questionnaire from me on different types of technology and services which are relevant to home consumers of telecommunication. This instrument is part of a Delphi study I am conducting in association with my dissertation research. This project is not officially associated with either the International Communication Association or the Human Communication Technology special interest group. I am writing again to thank those people who have alrehdy responded and to urge those who have not yet responded to please do so. Your responses are very important, as you know. With Maureen Beninson's help, the sample I selected for this study was to contain people most likely to be familiar with the concepts in the questionnaire. Several people have written back to me indicating that they wished to disqualify themselves from the panel because they did not feel comfortable with the content of the study. This is important information for me to have. Sumner is a difficult time to conduct a study, everyone seems to have a hectic schedule. But I would very much appreciate your help in completing this first round. As soon as the response rate is high enough, I can tabulate the Round One data and begin the Round Two process. If you have misplaced the questionnaire I have sent you, or if my first letter never reached you for some reason, please let me know. Thank you again, one and all, for taking time out of your busy schedules to help me. warm best regards, Richard V. Ducey (517)353-6410 APPENDIX C APPENDIX C INSTRUMENTATION FOR DELPHI STUDY - WAVE II MICHIGAN STATE UNIVERSITY acumen" Of TELECOMMUNICATION EAST LANSING ' MICHIGAN ' 0824 October 14, l982 Dear Fellow HCT Member: I would like to ask for your help in completing a Delphi study on telecommunica- tion technologies and services. I want to thank those of you who were able to participate in the first round of this study. I hope more people will be able to join us in this second round. I think this will be the final round because of :he overall consistency of estimates obtained in the first round of this stu y. As you know, the goal of a Delphi study is to generate a consistent or at least stable set of group-based judgements regarding some phenomena. To accomplish this, a respondent group is surveyed at least twice. The first survey is to obtain initial perceptions from the group. The second survey enlightens respon- dents to overall group opinion and provides the opportunity for individuals to make a second set of judgements based on a knowledge of how their peers have reacted in the first survey. In this second round, I have enclosed an instru- ment which is modified from the first round survey which does two things: (l) summarizes data from the first survey; and (2) asks you to make a second set of judgements in light of these data. The purpose of this study is to develop a reliable set of estimates regarding the extent to which various telecommunication technologies can provide a number‘ of teleservices. In this study, the focus of interest in the telecommunication link and the terminal device which would be used in providing the teleservices. I have enclosed a self-addressed and stamped return envelope. Please use this to return the survey instrument as soon as you have completed it. Hopefully, you will be able to complete this and have it in the mail before the end of this month. I am sincerely grateful for your help in this study. Thank you for contributing your time and expert knowledge to my project. Please call or write if you have any further questions. Harm best wishes, Richard V. Ducey (5l7) 353-64l0 RVDzaia Enclosures MSU is as Mira-stirs Adios/Equal Opportunity let-Wis. 1&39 190 Appendix C Directions for Round 11 Delphi Survey Section 1: Self Ratigg on Technologies 1. 2. Please rate your technical knowledge on the seven point scale indicated (l - not knowledgeable: 7 - knowledgeable). _'Technical Knowledge” means the degree to which you are familiar with the capabilities and limitations of each technology described for providing the kinds of services considered in this study. Section II: Technologies and Services I. 2. I have enclosed a description of services A-U for your reference. The task is to rate each telecommunication medium on each service using a seven point scale where: l . very inappropriate technolo ; 7 - very appropriate technology to support a given service. These judgements are subjective by nature, but should be grounded in considerations of technical attributes of the medium. Other considerations,. such as economic viability of providing different services via each medium may also influence your judgements. In this section, I have provided a measure of central tendency (mean) and dispersion (standard deviation) to summarize Round I data. Please indicate your estimate (using the 7 point scale described in (2) above) in the blank provided. Section III: Terminals l. 2. Section IV: Demographics Terminal design attributes are coded l-8. These are described on the survey instrument itself. The task is to indicate which feature or features (multiple choices allowed) the most appropriate terminal should have. "Appropriate" means features which would be needed for the service to be functional. One of the Round I respondents suggested another attribute - hard copy printing capability. If you think this attribute is needed for a given' service please circle the letter corresponding to that service in Column l of the table presented in’Section III. This section is fairly straightforward. 191 . Appendix C .}.;.'[ 7"“!1‘11 .‘uy October 1982 Round Two DIRECTIONS: This questionnaire is divided into several sections. The first section provides a few key word descriptors of each techndlogy in terms of (l) geographical service area, (2) bandwidth, (3) one-way vs. two-way and (4) type of conmunication voice. video, data. In this section you are asked to rate your own technical knowledge with respect to the technologies. In the next section you are asked to estimate the extent to which each telecomunication technology can support each service, and which type of terminal is most appropriate for each service. Finally, some demographic questions are asked to assist in the analysis of these data. . I. Self Ratigg on Technologies: For each technology described, please rate your. technicaliknowlédge on a seven point scale where: 7b knowledgeable and 1- unknowledgesble about this technology.- . not ' _ Knowledgeable Knowledgeable A. ONE-HAY CABLE TV I 2 3 4 5' 6 7 (limited service area--one or two towns; broadband; one-way only: voice, video and data). 8. THO-HAY CABLE TV ‘ l 2 3 4 5 6 7 (same as above. but data communication , - can be returned from the home,making it two-way) C. BROADCAST TV (UHF or VHF) l 2 3 4 5 6 7 (metro size service area; 6 MHz; ' . one-way; voice, video, data, one-way) o. BROADCAST RADIO (m or m) 1 2 3 4 5 6 7 ' (metro size to interstate service area; narrowband; voice. slow video; data; one-m) ' E.TELEPHONE l234'567 (virtually unlimited service area: - narrowband: voice, slow video, data; fully switched two-way, conference calls make it n-way) r. pas (direct broadcast satellite) l 2 3 4 "5 s 7 (service areas may correspond to U.S. time zones; voice, video. data; one-way) G. MOS (multipoint distribution service) l 2- 3~ 4 S 6 7 (service area limited to one city and ' inmediate surroundings; voice. video, data; one-way or two-way capable) H. LPTV (low power TV) l 2 3 4 S .6 7 (service area might extend to several communities; 6 MHz, voice, video, data, one-way) I. TERMINALS l 2 3 .4 S 6 7 (adiressabi l i ty, numerical keypad, alpha- numeric keypad, microprocessor equipped or not) Q .433 II. Technologies 8 Services R -2- Hgdtung One-way Cable!!! Your Service Mean S.D. Estimate” 1 1.342’ 1.893 I 2.632 2.166 C - g 1.895 1.997 D . 1.737 1.628 I 1.579 1.575 C 2.895 2.424 _______ I 1.056 0.236 I 1.895 . 1.997 J 1.3681 1.383 I 3.579 2. 755 L 4.053 2.549 a 1.947 2.121 I 1.611 1.787 O .. 6.056 1.924 P 1.667 1.782 ' 0 2.526 2.144 I 4.188 2.949. 3 2.000 2.171 I 1.211 0.918 0 1.222 0.943 Hediul: Broadcast TV _. Thur Service Mean S.D. Estimate A 1.000 0.000 3 1.789 1.437 C 1.000 0.000 D 1.000 0.000 8 1.38 1.243 r 1.83 lo 790 ' 0 2.57 2.219 ‘ I 1.000 0.000 - I 1.053 0.229 ‘ J 1.000 0.000 I 3.211 2.616 I 2.842 2.522 H. 1.000 ' 0.000 I 1.222 0.943 0 6.000 2.033 P 1.500 1.465 0 2.500 2.176 I 4.250 3.000 S 1.93 _ 2.120' I 1.000 0.000 0 1.000 0.000 192 - Medium: I‘m-wax Cable 11 Appendix C . ' . Your Service Mean ‘S.D. Estimate' A 6.100 1.586 8 6.150 1.309 C . 6.400 1.314 D 5.650 2.110 3 6.300 1.418 0 6.250 0.910 . a 6.158 1.015 I 6.050 1.432 J 5.850“ '1.872 - I 6.250 ' 0.910 L "6.400 1.231 a 6.200 1.508 H 4.842 2.433 0 ° 6.333 1.029 P 6.211 0.976 Q 5.750 1.650 I 6.000 1.372 8 6.294 1.160 T 4.700 2.473 0' 4.778 2.463 ' “gain-g Broadcast radio Your Service Mean S.D. Estimate A 1.000 0.000 I 1.526 1.307 C 1.000 0.000 0 1.000 0.000 ' 3 1.389 1.243 r 1.889 1.906 C 1.833 1.757 8 1.000 0.000 I 1.000 . 0.000 J 1.000' 0.000 I 2.316 2.405 L 1.000 0.000 a: 1.000 0.000 H 1.111 0.471 0 » 1.267 1.033 P 1.389 1.243 Q 1.500 1.339 3 3.938 2.955 3 1.800 2.111 T 1.000 0.000 0 1.000 0.000 II. (cont. Hedium: Service damndmomxrnunmnmuunub Hedium:v ) Telephone Service dflmwb'flOflKl‘HHD-‘IONNUOU’ Bean 5.900 4.900 5.200 ‘Hean 3. 263 3.739 3.368 3.‘21 3.722 3.222 5.000 3.339 2.395 2.342 4.000 3.421 2.395 1.722 5.250 3.333 3.339 a.313 3.313 2.153 1.395 1.447 1.997 1.824 2.231 2.038 2.208 2.646 2.038 1.209 0.923 2.328 2.007 0.813 2.588 2.324 11731 2.647 2.173 2.321 2.770 2.720 2.469 2.250 2.565 2.631 2.761 2.439 '2.494 2.547 2.514 2.544 2.603 2.673 2.492 1.776 2.380 2.744 2.570 2.536 2.798 2.167 2.079 Your 0 )1 Your Estimate 193 Appendix C Medium: D33 Your Service Mean S.D. Estimate A 1.105 0.315 8 1.421 0.961 C 1.105 0.315 o 1.053 0.229 I 1.278 0.826 P 1.833 1.724 C 2.316 2.136 a 2.056 1.862 I 1.368 1.383 J 1.316 1.376 I 2.789 2.463 L 3.579 2.714 14 1.421 1.305 8 1.000 0.000 O 5.875 1.996 P 1.944 2.182 O 2.667 2.326 8 4.250 3.000 S 2.333 2.469 ‘ T 1.316 1.376 0 1.333 1.414 Hsdiumc LPTV _ Your Service Hban S.D. Estimate‘ A 1.053 0.229 . 8 1.684 1.416 C 1.053 0.229 D 1.000 new 8 1.444 1.338 P 2.000 1.970 C 1.833 13505 a 1.111 0.471 . r ' I 1.053 0.229 - . r . J 1.000" 0.000 I 2.316 2.083 L 2.842 2.410 I! 1.211 0.713 I 1.278 1.179 O 4.688 2.651 P 10‘“ 1029‘ ‘ Q 2.222 i2.102 2 3.875 ‘ 3.008 ' ‘ 8 2.188 2.455 _ T 1.316 1.376 0 1.316 1.376 ‘ 194 Appendix C 111.: Terminals ‘4’ Termiggl Designs Attributes Selected (in x; n-ZO)* Service I 2 3 4 5 6 7 8 'a 88.2? _ $8.8 _ 11.8 _ 23.5 _ 5.9 __ 5.9 _ 17.6 _ 35.3 _ n 66.7 _ 88.9 __ 72.2 _ 77.8 _ 11.1 __ 5.6 _ 22.2 _ 11.1 _ c 87.5 _ 43.8 _ 6.3 _ 25.0 _ 6.3 _ 25.0 _ 12.5 _ 43.8 __ D 94.1 _ 58.8 _ 35.3 _ 47.1 _ 15.0 _ 11.8 _ 15.0 _ 23.5 _ s 82.4 _ 41.2 _ 52.9,_ 29.4— 15.0 _ 23.5 _ 58.8 _ 29.4 _ r 83.3 __ 83.3 _ 94.4 __ 83.3 __ 5.6 _, 5.6 __ 5.6 __ 0.0 _ o 33.3 _ 77.3 _ 94.1 _ 72.2 _ 5.0 _ 11.1 _ 11.1 _ 11.1 _ a 93.8 _ 87.5 _ 100.0 _ 93.8 __ 0.0 _ 0.0 __ 6.3 _ 0.0 _ I 88.9 _ 55.6 _ 55.6 _ 50.0 _ 0.0 __ 16.7 _ 50.0 _ 22.2 _ J 33.2 _ 64.7 _ 76.5 _ 53.3 _ 5.9 _ 11.3 _ 29.4 __ 11.3 _ x 73.5 _ 64.7 _ 33.2 _ 70.6 _ 11.3 _ 11.3 _ 11.3 _ 5.9 _ 1. 30.0 _ 25.7 _ 20.0 _ 33.3 _ 13.3 _ 30.0 _ 53.3 _ 40.0 _ 11 31.3 _ 50.0 _* 37.5 _ 43.3 _ 12.5 _ 25.0 __ 50.0 _ 31.3 _ II 68.8 __ 81.3 __ 93.8 __ 87.5 __ 12.5 __ 6.3 __ 12.5 __ 6.3 __ o 30.0 _ 30.0 __ 40.0 _ 20.0 _ 60.0 _ 60.0 _ 30.0 _ 60.0 _ P 66.7 __ 73.3 _ 30.0 _ 73.3 _ 20.0 __ 13.3 _ 20.0 _, 13.3 __ Q 64.7 _ 76.5 _ 76.5 _. 70.6 _ 17.6 _ 11.8 _ 23.5 __ 11.8 _ R 71.4 _ 78.6 _ 78.6 _ 85.7 _ 21.4 _ 14.3 _ 21.4 _ 7.1 _ s 66.7 _ 73.3_ 93.3_’_ 73.3__ 13.3_ 6.7_ 13.3_ 6.7 _ '1' 75.0 _ 87.5 _'_ 81.3 _ 93.8 _ 6.3 _ 6.3 _ 25.0 _ 0.0 _ 0 60.0 _ 86.7 _'__ 93.3 _ 86.7 _ 20.0-___ 0.0 _ 20.0 _ 0.0 _ *Due to some missing data, not all percentages are based on 20 responses. I DIRECTIONS: In this. table, percentages of people selecting each attribute for each service are indicated. In the space to the right of the data from Round One of this study," please place a check if you feel that a terminal providing the given service would need this attribute. In other words, for service C, 6.32 of the respondents in Round I indicated that attribute '13-(alphanume‘ric'keyboard) would be needed. 1: I agreed with this, 1 would filace a check in the blank next to the 6.3. Otherwise, I would leave that spot blank. '0001110 3*: _ .1- 'address'able ' 5- non-addressable - ‘ 2" microprocessor 0901PP¢d 6- non-microprocessor equipped 3" alphanumeric 3.75081“ 7- numeric keyboard (like typewriter) (like telephone with push buttons) 4" memory 8- no memory Thus, a combination of 1-2-7 describes a terminal which is addressable, bag a microprocessor and uses a numeric keypad. _ I t 2 III. 195 Appendix C -5- Demographics . A. How long have you been a member of HCT? O (months) 8. Hhat is your academic rank'or professional'title?‘ ° ( J . C., Please list the academic areas in which you have earned degrees. £312. ' ~ m. 8.A. H.A./".5. Ph.D. 4 Other 0. In addition to ICA, with what other professional associations do you affiliate? SCA. . _; AEJ IEEE MAPOR ACM OTHER (please specifiy) E. ‘33.: technically oriented journals, if any, do you read? (Please list) - F. :Age: . (years) 3.0 Gender: (1) Female - (2) Male 8. Did you participate in Round One of this Delphi survey? I o 9 YES 80 mm 100 AGAIN FOR YOUR rumor AND COOPERATIW III CONDUCTING THIS SURVEY!!! - I96 «apemao one: «go: so» eo.uascoe:. use «so coo an» on mucous; ac.nooxxooo gas» usage cu has pzmomao mov¢>oea .: ahouzaeou a a? mcwaoox -xooa u—ocowaog can» ea p—o coax ou.»u__.nanau .p ~mum=oamoh Oaxaca opoevm mo coupon. momMmmos ppse mama so» pop opaos gu.ga veooaaux asp—.gou.cawaxu a use: on au._—aaaau .m ~z<¢pmoui to u~mag no: we. moo—>som omogu on: o» xu——.naaou .a «mes—noun gums soc» mu—amoe ceaumcouea so» n—o: o» manage to acumen no gush massage; ouvsosa cu xu—__aaa~u .a ~u. ago: an» ou_uov so» .eoueo.A—ouuvuossw eo—uaEcome— you on zap—vaunuu .a ~uum covm—>opou sac» co copou e. cream on mocauu.a tea mono: 0:» mo ppa u>a= o» xu——.aaaau .o ~xsosweog ocean no .moxau 9:3 .3055... 32.032. mo 3.095 9.309. 9.: mEopaohn game on so» a—u: space so—gr ou—>som < .z ~ueo: each a. aueoonueo _au—voe a .ocm.m appau—uaeouaa no» no. apnea =u_=r ou—sson c .2 NOUOL 069% 0 96-369 22.-H LOEHG& aUOSUOOIJJ. 2238 do» 25 8:8 82: he :5 as. an» amp u—so: ooze ou.>aom o_>os eu.>.aoauaoa < .4 .sPQEaxo toe .ocaummzue a on maven . an new a__o:ma e.gou so» eovuaELome_ eo cap; as» on v.39: m_gh anon ao.n.>opou each so moaauu—a was «who: no green on e.gou gaps: co.uoesome— someone. panacea you so» no. open: nag» ouv>som < .2 Appendix C «mason sac» eon» a_—au. -uaeouaa m__.a and an» uo— u—aos was» ou.>eom < ~mu==ooua xaoa s:o».eo hues—an as» «so 2.: 3.: 355 8 so» so. use, as: 8.28 e «and: seam.e— ou.>aom p.ae capauwa one use, a. a. one?“ ..os upcoeuuopo m—gu ppau spaced neon «swoon ween: mo unname— uom covm.> .owwu sac» co crash on opaou gu_:s s_aooa ahead on no chaos that no echo :9» no. space one» wov>eon < has: a: 3 9:8 no.uwa afipasnoe agave so» eo.uasaou:. we we.“ see a. mesh «cannon eo—u—so—ou can» go manage. u:- mutoa ma green on vpaoo no.3: eovuaELoee. unocuue. pa—uwam «on :oa.uw—.s—:oa «as» oo_>cwu < sues: can» ease appau—aaeouaa newsman seen so nee—umoaa an. c—aou so» other .gmra aoa.e— .u_uoeu was—poo soc woueaou —oo=um use» aoa.uo_ spas: ease ou_>eoa < . pup—o: covepao. so meo.uuopo pauc— e. use: ease sues aca_uo— specs nag“ ou.>csa < tacos . s>aa a». house e.. .vwoowe ace who use» use: «use: can» :— u:.eo.u.veoo.c.a so can; we» can Loans: scans «on us» grow ens» appaueuaeouaa a» use: a:oa.ou touaaeou a uuueecu amen: uncu ouv>eau c puswaae xa.p.u= eaoa.uaoa upped—unsouaa cu use: eaoa.ou sausages a uuoaeou v—aoe nag» quests» < ~85: .52. sec mason 83> pm. woe... estate... 3.. ..e. no. 2:2. 5:... 8.3.8 .1 pasta-- .3323 a... 2: 858.33 03.6.... 3 85... . taoa_oa aaaaaeou a uuweeoo space «age ou~>eom < 197 Appendix C MICHIGAN STATE UNIVERSITY DEPARTMENT or WUNCATIGU EAST LANSING ' MICHIGAN ' 48824 November l6, 1982 Dear Fellow HCT Member: About three weeks ago you should have received a questionnaire from me on different types of technologies and services which are relevant to home consumers of telecommunication. This instrument constitutes the second round of a Delphi Study which I am conducting as part of my dissertation research. This project is not officially associated with either the International Communication Association or the Human Communication Special Interest Group. I am writing again to thank those people who have already responded and to urge those who have not yet had a chance to respond, to please do so. As you know, all responses are very important. Even if you did not participate in Round One of this Delphi Study, please feel free to participate in Round Two. I very much appreciate all of your assistance and patience with this project. If you have misplaced the questionnaire I sent you, or if you have any other questions please call or write to me. Many thanks for your support. Harm best regards, Richard V. Ducey (5l7) 353-6410 RVD:aia MSU is - Allan-slur Acton/Equal Wv Imam-nos