| Home | Audio Magazine | Stereo Review magazine | Good Sound | Troubleshooting |
|
Influenced in part by these anthropologists, several empirically minded sociologists of the 1920's and 1930's also demonstrated an interest in diffusion. Studies were made of the spread of the city-manager plan, of a third-party movement, of amateur radio as a hobby, and the like. The guiding theoretical concerns had to do with the influence of the metropolis on its satellites, the effectiveness of natural and legal boundaries as barriers to diffusion, the flow of innovation from region to region across the country, as well as the hypothesis of a "concentric circle" pattern of diffusion which was shared with the anthropologists. The underlying assumption was always that informal communication among adopters was the key to diffusion. In both of these fields, diffusion studies came to a halt by about 1940. In anthropology, attention shifted to the closely related problem of acculturation5 in which emphasis is placed on ongoing (rather than historical) situations of intergroup contact, on patterns of culture traits rather than single items and, typically, on pairs of interacting societies rather than longer chains of connected groups. It is less clear why diffusion studies failed to hold the interest of sociologists, though they were never as prevalent as in anthropology. It seems a reason able guess, however, that the revolution in communication which began with the rapid spread of radio in the late 1920's and early 1930's diverted their attention. THE REVIVAL OF INTEREST IN DIFFUSION RESEARCH But the mass media are incapable of influencing people (though they may inform them) as directly or as simultaneously as had been imagined. Indeed, the study of mass media "effects," with its primarily psychological bias, is now broadening to take account of the social processes involved in the spread of influence and innovation. This seems an altogether reasonable next step for former students of mass media "campaigns." For, if the mass media are not as all-powerful as was originally imagined, the problem of understanding the furious rate at which new ideas and behavior travel through society still remains. In short, there is a revival of interest in diffusion processes. The sociologists of communication who found themselves interested in dif fusion discovered, somewhat to their surprise, that relevant studies were being carried on in a number of closely related fields. The most conspicuous case is that of rural sociology which has accumulated, over the last two decades, several hundred studies of the communication and acceptance of new farm practices. Similarly, researchers in the field of education have tried to understand the rate of acceptance of innovations by school systems and have looked at such things as the spread of the kindergarten or supplementary reading. Public health is interested in the acceptance of new health practices-the Salk vaccine, for example." Marketing researchers, of course, are interested in the spread of acceptance of new products (although they have done far less work on this problem than one might imagine); folklorists have documented the extent to which children's games, for example, have spread from region to region; " and so on. ----------------------- From Elihu Katz, Martin L. Levin, and Herbert Hamilton, "Traditions of Research on the Diffusion of Innovation," American Sociological Review, April, 1963, 28, 237-252. Reproduced with permission of the authors and publisher. Preliminary formulations of portions of this paper were presented at the meetings of the American Sociological Association, Chicago, 1959, and the American Anthropological Association, Mexico City, 1959. We wish to thank Robert L. CraM and Manning Nash for reading and commenting on the present version; readers of the earlier version are too numerous to mention. The overall project of which this paper is a part has received the support of the Foundation for Research on Human Behavior and the Social Science Research Committee of the University of Chicago. 1. Robert H. Lowie, History of Ethnological Theory contains an excellent treatment of the early anthropological movements and schools, including evolutionism and diffusion. In this connection also see Alexander Goldenweiser's "Cultural Anthropology" in Harry Elmer Barnes (ed.), History and Prospects of the Social Sciences, New York: Alfred A. Knopf, 1925. A. L. Kroeber's article "Diffusionism" in the Encyclopedia of the Social Sciences is a brief, interesting description of the early diffusionist work in the context of the development of anthropology, and Melville Herskovits, Man and His Works, Chapters 30 and 31, New York: Alfred A. Knopf, 1938, can also be examined with profit as an informative survey of the various early movements in anthropology. For various sides of the argument concerning the early work on diffusion see G. Elliot Smith, et al, Culture: The Diffusion Controversy, New York: Norton, 1927. 2. Among the more suggestive of these studies for our purposes are the following: Robert H. Lowie, "Plains Indians Age Societies," Anthropological Papers, American Museum of Natural History, 11 (1916), pp. 877-1031; Robert H. Lowie, "Ceremonialism in North America," American Anthropologist, 16 (October-December, 1914), pp. 602-631; Paul Radin, "A Sketch of the Peyote Cult of the Winnebago: A Study in Borrowing," Journal of Religious Psychology, 7 (January, 1914), pp. 1-22; Leslie Spier, "The Sun Dance of the Plains Indians: Its Development and Diffusion," Anthropological Papers, American Museum of Natural History, XVI (1921), pp. 451-527; Clark Wissler, "Material Culture of the North American Indians," American Anthropologist, 16 (October-December, 1914), pp. 477-505; Clark Wissler, "Costumes of the Plains Indians," Anthropological Papers, American Museum of Natural History, 17 (1915), pp. 39-91; Clark Wissler, "The Ceremonial Bundles of the Blackfoot Indians," Anthropological Papers, American Museum of Natural History, VII (1914), pp. 65-289. 3. See, for example, Raymond V. Bowers, "The Direction of Intra-Societal Diffusion," American Sociological Review, 2 (December, 1937), pp. 826-836; F. S. Chapin, Cultural Change, New York: Century, 1928; Edgar C. McVoy, "Patterns of Diffusion in the United States," American Sociological Review, 5 (April, 1940), pp. 219-227; H. Earl Pemberton, "Culture Diffusion Gradients," American Journal of Sociology, 42 (September, 1936), pp. 226-233. 4. In the 1920's and 1930's, a method of diffusion research commonly referred to as "age-area" analysis was developed. This method involved reconstructing the temporal movement and spread of cultural traits and complexes from geographic data on the spatial distribution of the cultural elements under investigation. Especially noteworthy in this regard was the work of Clark Wissler, who developed it to its most refined degree. See his Man and Culture, New York: Thomas Y. Crowell Co., 1923, and The Relation of Nature to Man in Aboriginal America, New York: Oxford University Press, 1926. While a very important contribution to the field in its day, this approach was subjected to a searching critique by Roland Dixon in The Building of Culture, New York: Scribners, 1928. This kind of criticism, no doubt, contributed appreciably to the subsequent decline of distributional diffusion studies generally. Still, there has been a continued production of such studies and, indeed, one occasionally encounters especially interesting investigations at least partly employing this approach, such as the recent work by David F. Aberle and Orner C. Stewart, "Navaho and Ute Peyotism: A Chronological and Distributional Study," University of Colorado Studies, Series in Anthropology No. 6, 1957. Furthermore, those interested in archaeology seem to have maintained an even more central concern for diffusion analysis along these distributional lines. See, for example, the spirited discussion of the paper by Munro S. Edmondson, "Neolithic Diffusion Rates," Current Anthropology, 2 (1961), pp. 71-102. 5. Acculturation or culture contact studies were heralded by Robert Redfield, Ralph Linton and Melville J. Herskovits, in their "Memorandum on the Study of Acculturation," American Anthropologist, 38 (January-March, 1936), pp. 149-152. 6. Indeed, a large segment of diffusion studies has tended to concern itself with adjustive responses to contact rather than the transmission of items between groups. This paper will explicitly avoid consideration of the now predominant concern with the social and cultural consequences of change. Our focus is on the processes of communication of change. Studies concerned with non-diffusion aspects of change are helpful, however, in drawing attention to the interrelationships among diffusion processes, socialization processes and adjustive processes in culture change. For an interesting empirical study illustrative of the link be tween socialization processes and (resistance to) acculturation, see Edward M. Bruner, "Primary Group Experience and the Processes of Acculturation," American Anthropologist, 58 (August, 1956), pp. 605-623. 7. For a discussion of some of the social and psychological factors involved in the trans mission of influence via the mass media, see Joseph T. Klapper, The Effects of Mass Communication, New York: Free Press, 1960, Part One. ----------------------- ----------------------- 8. The design of research in mass communication has recently begun to take account of interpersonal relations as structures which relay and reinforce (or block) the flow of in fluence and innovation. See Elihu Katz, "The Two-Step Flow of Communication: An Up-to Date Report on an Hypothesis," Public Opinion Quarterly, 21 (Spring, 1957), pp. 61-78. 9. In the few relevant studies so far, the tendency has been to follow a communication as it passes from one individual to the next, to establish the nature of the relationship between the interacting individuals and thus to infer the relevant social networks; in other words, structures of social relations are derived from the flow of interpersonal communication. The alternative method-that of mapping the potentially relevant structures of social relations prior to tracing the flow of influence-would seem to be somewhat more desirable, if more difficult. 10. For an overview of work in this field together with selected bibliography, see Herbert F. Lionherger, Adoption of New Ideas and Practices, Ames, Iowa: Iowa State University Press, 1960. 11 See Paul R. Mort and Frances G. Cornell. American Schools in Transition, New York: Teachers College, Columbia University, 1941; and Walter Cocking, "The Regional Introduction of Educational Practices," New York: Bureau of Publications, Teachers College, Columbia University, 1951. 12. For numerous references, see Steven Polgar, "Health and Human Behavior: Areas of Interest Common to Social and Medical Sciences," Current Anthropology, 3 (April, 1962), pp. 159-179, particularly the section on Health Action Programs and, to a certain extent, the section on Dynamics of Health Status. Anthropologists have been particularly active in this area. 13. The most interesting study, from our point of view, is The Tastemakers (Vol. I), a report of the Public Opinion Index for Industry, Princeton, New Jersey: Opinion Research Corporation, April 1959. 14. See, for example, Iona and Peter Opie, The Lore and Language of Schoolchildren, London: Oxford University Press, 1959. ------------------------------- Like sociology, anthropology has also experienced something of a return to some of the interests of the more sober schools of diffusion, as a by-product of the current effort to evaluate the progress of the varied programs for planned change in underdeveloped areas of the world. THE STATE OF DIFFUSION RESEARCH Ironically, it almost seems as if diffusion research in the various research traditions can be said to have been "independently invented!" Indeed, diffusion re searchers in the several traditions which we have examined scarcely know of each other's existence. The recent "discovery" of rural sociology by students of mass communications and vice versa is a good case in point. As a result, each tradition has emphasized rather different variables and a characteristically different approach. This paper attempts to integrate these diverse points of view. To accomplish this, we shall first propose a working conception of diffusion from a sociological point of view. This will be done in terms of a tentative set of component elements, each of which can be formulated as a key variable (some times as several variables) intrinsic to, or bearing upon, the diffusion process. Taken together, they constitute a kind of "accounting scheme" for the study of diffusion. Following the enumeration of the component elements, each will be considered in some detail, paying particular attention to problems of conceptualization and operational definition. Then, we shall attempt to "locate" the characteristic emphases of each of the research traditions in terms of one or more of these elements of the diffusion process. DEFINING DIFFUSION Viewed sociologically, the process of diffusion may be characterized as the (1) acceptance, (2) over time, (3) of some specific item-an idea or practice, (4) by individuals, groups or other adopting units, linked (5) to specific channels of communication, (6) to a social structure, and (7) to a given system of values, or culture. ---------------- 15. There is a burgeoning literature on this subject. See the studies reported in recent volumes of Human Organization and Economic Development and Cultural Change; the several collections of case studies, particularly Benjamin Paul (ed.), Health, Culture and Com munity, New York: Russell Sage Foundation, 1955; and recent volumes such as Charles Erasmus, Man Takes Control, Minneapolis: University of Minnesota Press, 1961; and George M. Foster, Traditional Cultures and the impact of Technological Change, New York: Harper, 1962. 16. For an account of the confrontation between students of mass communication and of rural sociology, see Elihu Katz, "Communication Research and the Image of Society: Convergence of Two Traditions," American Journal of Sociology, 65 (March, 1960), pp. 435-440. --------------------- Altogether, there are very few studies in any of the traditions of diffusion re search which have incorporated all of these elements. In fact, the traditions differ from each other precisely in their tendency to "favor" certain of the elements rather than others. Now we shall consider each of the components in turn. 1. ACCEPTANCE. Acceptance is the dependent variable in most studies of diffusion though, strictly speaking it is time-of-acceptance that is really of interest. Ideally, in other words, diffusion studies seek to classify acceptors in terms of the timing of their acceptance of an item or to compare the relative rate of acceptance in one community with another. More often than not, however, information about time is lacking and, instead, one learns-for a given point in time-which individuals have and have not accepted an innovation or what proportion of com munity members in different communities have accepted. Most diffusion studies define acceptance rather arbitrarily. Where information on time is available, date of "first use" is frequently employed as the measure of acceptance, the season of first-use of hybrid corn, for example." But, obviously, first-use may or may not be followed by continued use and some recent studies, therefore, have insisted on the distinction between "trial" and "adoption." Thus, a measure of "sustained use" might be appropriate for some purposes but, for other purposes, it may be of interest to consider only "ever use." For anthropologists, however, this is a much more serious matter. First of all, anthropologists tend to be skeptical about the extent to which a given item is perceived and used in the same manner in different societies. If the sewing machine is prominently displayed on the open porch, but never used for sewing, it may be argued that it is no longer the "same" item. For anthropologists, that is, acceptance tends to refer not to the form of an item alone but to form-mean ing-function.1 ° Consider the acceptance of Christianity, for example, as discussed in the anthropological literature. With respect to its appearance in a given society, anthropologists would tend to ask: (1) Is it the "same" item? (2) Is it internalized in the personalities of the group? (3) Is it central to the social institutions of the group? Indeed, one of the factors underlying the distinction between "acculturation" (a prestigeful concept) and "diffusion" (a less prestigeful one) in anthropology, appears to be related to the "level" of acceptance (internalization and centrality) involved." This is a good example, perhaps, of the utility of confronting several traditions with each other within a manageable framework. Obviously, some kind of distinction must be made between mere external acceptance of a form and its internalization; and, obviously, attention must be given to the extent to which function travels together with form. But these ideas should not be treated merely as cautionary; they are also suggestive of hypotheses. Indeed, writing in a very similar vein, Gabriel Tarde-the social theorist of diffusion par excellence--suggested that "inner" changes precede "outer" changes in the sense that the diffusion of an idea precedes the diffusion of the tangible manifestation of that idea or, in other words, that there is a "material lag" rather than a "cultural lag" in the transfer of items across societal boundaries. Some theorists would agree; others, obviously, would not. In any case, the implication is that diffusion re search ought not to be misled by the argument over whether "mere diffusion" or penetration to deeper levels is more important but, rather, whether these correspond to separable episodes in the spread of any given item and, if so, how they are related. -------------------------- 17. See Bryce Ryan and Neal Gross, "The Diffusion of Hybrid Seed Corn in Two Iowa Communities," Rural Sociology, 7 (March, 1943), pp. 15-24. 18. Gaining acceptance for most contraceptive techniques, for example, is much more a problem of "sustained use" than of "first use." See Reuben Hill, J. Mayone Stycos and Kurt Back, The Family and Population Control, Chapel Hill: University of North Carolina Press, 1959. Obviously, the distinction is appropriate wherever first use does not lead directly to continued use. See A. Apodaca, "Corn and Custom," in E. H. Spicer (ed.), Human Problems in Technological Change, New York: Russell Sage Foundation, 1952, for a study of the acceptance of an innovation which was later discontinued. 18. The form-meaning-function distinction is stressed particularly in Ralph Linton, The Study of Man, New York: Appleton-Century, 1936, pp. 402-404. ------------------------ 2. TIME. If any one of the elements may be said to be more characteristic of the diffusion process than the others, it is time. It is the element of time that differentiates the study of diffusion both from the study of mass communication "campaigns" with their assumed immediacy of impact and from traditional distributional studies. Diffusion takes time; for example, it took ten years for hybrid corn-an unusually successful innovation-to reach near-complete acceptance in Iowa communities. Nevertheless, there are very few studies, so far, that have taken systematic account of time in the study of diffusion. In part, this neglect is a result of the difficulty of obtaining data. Studies which have taken account of time have relied on one of the following three methods: recall (where a respondent, or an informant, dates the acceptance of an innovation), records (where time-of-acceptance is a matter of record, for some reason), and inferences (such as in the archaeological dating methods of stratigraphy or Carbon-14). Some of the early diffusion studies by sociologists had access to data on time because they studied innovations intended for adoption by municipalities-the city manager scheme, for example. A current study of the diffusion of fluoridation has such data for the same reason. The dates of acceptance of such innovations are a matter of public record. ----------------------- For good examples of the applicability of the notion of levels of incorporation of an innovation into a receiving society, see Edward P. Dozier, "Forced and Permissive Acculturation," American Indian, 7 (Spring, 1955), pp. 38-44; and Edward H. Spicer, "Spanish-Indian Acculturation in the Southwest," American Anthropologist, 56 (August, 1954), pp. 663-678. 21. In his analysis of the sun dance, Spier explicitly cites his data as evidence, at least in this case against Tarde. See Leslie Spier, op. cit., p. 501. 22 See, for example, McVoy, op. cit. ------------------------------ Anthropologists who were studying diffusion in the 1920's and 1930's gave considerable thought to the development of a methodology for inferring time from spatial distributions. Clark Wissler, for example, was able to demonstrate that a particular distribution of pottery around a hypothesized point of origin did, in deed, correspond to a known succession of types of pottery as established stratigraphically. Wissler further indicated that "students of culture generally assume that widely distributed trait complexes are the older," though he immediately cautions that such an assumption may result in serious error insofar as the rates of diffusion of different sorts of items may vary. Early sociological students of diffusion faced a similar obstacle though they had more data on time. It is relatively easy to establish, for example, the date on which 10 percent of the population of a city or state owned a refrigerator or a radio. Then, treating the city or the state as if they were "adopters" of refrigerators and radios makes it tempting to suggest that certain cities are influencing others to adopt or that there seems to be a certain kind of geographical movement from state to state. A genuinely pioneering (though perhaps unconvincing) effort to strengthen this sort of tenuous ecological analysis with data gathered from individuals was made by Bowers in the 1930's. Bowers studied the diffusion of amateur radio as a hobby and demonstrated, for example, that the proportion of amateurs to population was at its highest in 1914-1915 in cities of 25,000-100,000; five years later the peak was in cities of 10,000-25,000; during the following five years, the heaviest concentrations were in still smaller cities. ----------------------- 23 See Robert L. Crain, "Inter-City Influence in the Diffusion of Fluoridation," unpublished Ph.D. dissertation, Department of Sociology, University of Chicago, 1962. 24 See Man and Culture, op. cit. Also see the work of Margaret T. Hodgen, "Geographical Diffusion as a Criterion of Age," American Anthropologist, 44 (1942), pp. 345-368. The article by Edmondson, op. cit., is based on Carbon-14 datings. 25 See Bowers, op. cit. -------------------------- From this distribution, he infers that people in the larger cities had influenced those in smaller places. Then, by means of a mail questionnaire, an attempt was made to test this inference by asking licensed amateur radio operators to report on the sources which were influential in their decisions to become "hams." If students of pre-history sorely felt the lack of data on time, it is a nice anomaly that students of "consumer" innovations in the mid-twentieth century are experiencing the same problem. It may be possible to ask a farmer to try to recall the season during which he first planted hybrid corn, but it is very difficult to be certain that such information is reliable. How much more is this the case for innovations which are less central to their adopters and inherently less datable than is the season of first use of a new kind of seed. One can perhaps ask about the date of purchase of major appliances, but it is almost impossible to rely on recall for most other things. A promising source for data of this kind is the type of consumer panel in which households are asked to keep a record of all their purchases, entering them in some sort of log on a daily or weekly basis; however, there are many difficulties with this procedure. Occasionally, unusual opportunities present themselves for obtaining data on adoption dates. The study of the diffusion of new drugs among physicians, for example, had access to prescriptions on file in local pharmacies, making it possible to date each doctor's first use of a new drug. Time is a crucial ingredient in the diffusion process, however, not simply because it enables the researcher to identify the characteristics of early-adopting individuals or to establish the direction of the flow of influence. It is also important because it provides a basis for the charting of diffusion curves, thus making possible the development of mathematical descriptions of variations in the diffusion process. Time, and the number of adopters at a given time, are continuous and easily quantified variables; hence, the study of diffusion is one of the areas of social science which lends itself immediately to the construction of mathematical models. For example, one can construct theoretical models of the diffusion process given certain assumptions and compare the results with those actually observed in the real world. On the basis of such a comparison, one can infer whether a given item is "contagious" or not, that is, whether the item spread as a function of the extra adoptions or the character of contacts with previous adopters. Hagerstrand, a geographer, was able to demonstrate that the most probable adopter of a new farm practice is the farmer living in the vicinity of someone who has just adopted it; and on the macro-level an innovation spreads from primary centers until the original source of influence is exhausted, whereupon some new center springs up. Crain found essentially the same phenomenon at work in the case of fluoridation where the unit of adoption is a municipality rather than an individual." Similarly, attempts have been made to specify, a priori, the probable influence of different patterns of social relations on the spread of innovation. The work reported by Stuart Dodd is a good ex ample."' The same kind of logic suggests that similar innovations may be de scribed by similar curves of diffusion and, if this is so, part of the problem of classifying innovations (to be discussed below) will be open to solution." ----------------------- 26. There is considerable difficulty in maintaining the representativeness of the consumer panel sample, and constant programmed turnover is one of the strategies of doing so; for diffusion research, however, turnover represents a complication. Moreover, if one approaches the diffusion problem in a situation where individual adopters are widely dispersed-such as in a national study of some consumer innovation, for example-one must cope with the added complexity of differing beginning dates in different regions, etc. 27. See James S. Coleman, Elihu Katz and Herbert Menzel, "The Diffusion of an Innovation among Physicians," Sociometry, 20 (December, 1957), pp. 253-270. 28 An impressive amount of work is going on in this area, much of it beyond the competence of the present authors. The major contributions include the following: Stuart C. Dodd, "Diffusion Is Predictable: Testing Probability Models for Laws of Interaction," American Sociological Review, 20 (August, 1955), 392-401; and Stuart C. Dodd and Marilyn McCurtain, "The Logistic Law in Communication," in National Institute of Social and Behavioral Science, Symposia Studies Series No. 8, Series Research in Social Psychology, Washington, D.C., 1961; Melvin DeFleur and Otto Larsen, The Flow of Information, New York: Harper, 1958; Georg Karlsson, Social Mechanisms, Glencoe, Ill.: The Free Press, 1958; Torsten Hagerstrand, "Monte Carlo Simulation of Diffusion," University of Lund, Sweden, 1960 (unpublished) ; and James S. Coleman, "Diffusion in Incomplete Social Structures," Baltimore: Department of Social Relations, Johns Hopkins University, 1961 (unpublished). Two economists who have worked intensively with diffusion curves are Zvi Griliches, "Hybrid Corn: Explorations in the Economics of Technological Change," Econometrica, 25 (October, 1957), pp. 501-522; and Edwin Mansfield, "Technical Change and the Rate of Imitation," ----------------------------- 3. A SPECIFIC ITEM. The discussion of acceptance has already made clear part of the problem of specifying the particular item under study. Obviously, one would like to ascertain whether the meaning of a given item for one individual, or for one society, is the same as it is for another. In a related sense, one would also like to know whether or not a given item is part of a larger "complex" of items to which it adheres. On the other hand, this does not preclude-as some people seem to think-the legitimacy of studying the diffusion of an isolated item, concentrating on form alone regardless of possible "adhesions" and regardless of possible variations in function. In any event, these problems are somewhat reduced when the items involved are practices more than ideas, items of lesser rather than greater pervasiveness, and when the study is concentrating on diffusion within a particular culture rather than across cultures. This, perhaps, makes somewhat clearer why anthropologists, more than others, have raised questions in this area. The major problem of specifying the item in diffusion research derives from these considerations. It is the problem of how to classify items so that the results obtained are generalizable to other items. This problem is not unique to diffusion research, of course, but it is perhaps particularly obvious in this context. Suppose one studies the diffusion of hybrid corn, or of fluoridation, or of 2-4-D weed spray. Unless some scheme of classification exists which would make it possible to say that a given new item is rather more like a 2-4-D weed spray than it is like hybrid corn, each study simply becomes a discrete case which cannot be generalized. -------------------- Pittsburgh: Graduate School of Industrial Administration, Carnegie Institute of Technology (unpublished). See also a noteworthy series of articles by bio-physicist Anatol Rapoport entitled, "Spread of Information through a Population with Socio-Structural Bias," Bulletin of Mathematical Biophysics, 15 and 16 (1953-54). Related work in the epidemiology of contagious disease is that of N. T. J. Bailey, The Mathematical Theory of Epidemics, London: Charles Griffin, 1957. Steven Polgar has written a paper that is relevant here on "The Convergence of Epidemiology and Anthropology," School of Public Health, University of California at Berkeley (unpublished). 29 Hagerstrand, op. cit. 30. Crain, op. cit. 31. Dodd and McCurtain, op. cit. 32. This is more difficult than it sounds, perhaps, but it is a lead worth following. ---------------------------- Such a classification system is particularly difficult because, like all "con tent analysis," one must make guesses about the meaning of the item to its potential audience. Of course, to a certain extent this can be studied empirically. Suppose, for example, that the dimension of "radicalness"-that is, the extent to which an innovation is a major departure from some previous mode of acting- were an important one, as many observers seem to think. One might pre-test the actual use of an innovation--a visual telephone, for example--to discover the kinds of behavioral and attitudinal changes which it implies in order to rank it, at least as perceived by its early users, on a radicalness scale. But the trouble is that nobody is quite sure what dimensions of an item are relevant, and very little research has been done to try to find out. There are some exceptions, however. Wilkening in the United States and Emery and Oeser in Australia have traced the spread of several different agricultural innovations through the same communities and, on the basis of their differential rates and patterns of acceptance have speculated about some of the dimensions which affect diffusion." A major study of the diffusion of educational practices also speculates about why different sorts of innovations seem to spread in different patterns."4 Dimensions that have been suggested by these authors and others center on economic-sounding considerations such as (1) extent of capital outlay required; (2) extent of anticipated profitability; (3) certainty of profitability or efficacy, and extent of possible loss or danger (risk). Of course, these are not strictly financial matters at all. The most promising works on this problem have been several attempts to explicate the most traditional of the dimensions in terms of which innovations have been classified: material vs. non-material items. Barnett and others have suggested that material items find more ready acceptance because (1) they are more easily communicated; (2) their utility is more readily demonstrable; and (3) typically, they are perceived as having fewer ramifications in other spheres of personal and social life. Following Barnett, Menzel classified several different kinds of medical innovations in terms of his estimates of their (1) communicability, (2) risk, and (3) pervasiveness, hypothesizing that early adopters of each item would have certain characteristics. He suggested, for example, that integration into the local medical community would be characteristic of early adopters of a new drug which "required" communication but neither risk nor pervasiveness, whereas acceptance of a psychotherapeutic technique would be likely to "require" a certain emancipation from the local community and thus lesser integration. The results obtained were promising and represent the opening up of an important direction for diffusion research. -------------------------- 33. Eugene A. Wilkening, Acceptance of Improved Farm Practices in Three Coastal Plains Communities, Raleigh: North Carolina Agricultural Experiment Station, Bulletin 98, 1952; and F. E. Emery, Oscar Oeser and Joan Tully, Information Decision and Action: A Study of the Psychological Determinants of Changes in Farming Techniques, Carleton: Melbourne University Press, 1958. 34. Mort and Cornell, op. cit. 35. Barnett, op. cit., pp. 374-377. 36. Herbert Menzel, "Innovation, Integration and Marginality," American Sociological Review, 25 (October, 1960), pp. 704-713. ------------------------ 4. UNITS OF ADOPTION. Another way in which items can be usefully classified is in terms of the units of adoption for which they are intended. Most studies in sociology, rural sociology and marketing have considered only consumer-type items, those intended for adoption by an individual. But some innovations are intended for-indeed, they may "require"-groups, in the sense that it "takes two to tango" (or to telephone, or to perform the peyote ritual, etc.). And among such group-oriented innovations, a further distinction seems useful. There are items which require collective adoptions but permit any given individ ual to adopt or not (the telephone, for example) ; there are other items, however, where the group adopts as a single unit leaving no room for individual options I fluoridation, for example). Just as the item may "require" one or another adopting unit, a given culture may "prescribe" one rather than another adopting unit as appropriate. The kibbutz prescribes a group decision even for consumer-type innovations intended for use by individuals; similarly, the simultaneous conversion of an entire village to Christianity reflects the acceptance of a corporate decision, made by the chief perhaps, as binding upon all. Anthropologists are much more likely than those in other traditions to focus on the group as an adopting unit. Sometimes, this is just another way of talking about individuals as, for example, when it is reported that Village A adopted a certain kind of plow but Village B did not. But, often, the group is indeed the unit of adoption in the sense that the group "decides," or the culture "prescribes," that there be a collective decision. Thus, the unit of adoption may vary as a function of the "requirements" of the item or the "prescription" of the culture. And, just as in the case of the other elements in the diffusion process, the adopting unit functions as a variable to facilitate or block the flow of acceptance of innovation. For example, items which "require" collective adopting units may be resisted, therefore, by cultures which "prescribe," or favor, the individual as the unit of adoption and vice versa. Resistance to fluoridation, in the United States, in terms of minority rights is one such example; resistance to consumer innovations by Israeli kibbutzim is another. By the same token, an appeal for acceptance of an innovation is less likely of success when directed to the "wrong" adopting unit-as when family planning campaigns aim at, say, the wife, but the culture "prescribes," or the technique "requires," joint agreement by both spouses." ----------------- 37. For further discussion of the points raised in this section, see Elihu Katz, "Notes on the Unit of Adoption in Diffusion Research," Sociological Inquiry, 32 (1962), pp. 3-9. ----------------- 5. CHANNELS. So far, almost nothing has been said about the channels which transmit information and influence concerning an innovation. Indeed, except for occasional studies which noted the role of highways or of caravan routes, channels-like time-are missing in most of the early studies of anthropologists and sociologists. Even when it seemed certain, from distributional evidence or other inferences, that an innovation traveled from Tribe A to Tribe B, it was often unclear how this took place. On the other hand, if there is any single thing that is most wrong with contemporary studies of diffusion in the fields of mass communication, rural sociology and marketing research, it is that there is too much emphasis on channels. The typical design for research in these fields has been based, almost exclusively, on the assumption that people can be asked to recall the channels of information and influence that went into the making of their decisions to adopt an innovation or to make some sort of behavioral change. This approach in mass media research is known as "reconstruction" or "reason analysis." It is of some methodological interest, too, because it reverses the usual experimental design of "campaign" studies which begin with stimuli and try to track down their effects. Reason analysis, instead, begins with an effect and seeks to reconstruct how it came about. It is this approach which is, in a sense, responsible for the rediscovery of the importance of inter personal relations in the flow of influence and innovation in modern society. It is only very recently that students of mass communications and marketing have begun to include interpersonal relations among the channels of diffusion. This contrasts sharply with the rural sociologists who have long been aware-though they have not formulated it systematically until rather recently-that there is a "two-step flow" from the county agent to an influential farmer and thence to other farmers. ------------------ 38. For discussions of "reason analysis," see Paul F. Lazarsfeld and Morris Rosenberg (eds.), The Language of Social Research, Glencoe: The Free Press, 1955; and Hans Zeisel, Say It with Figures, New York: Harper, 1957. 39. Clark Wissler's "The Influence of the Horse in the Development of Plains Culture," American Anthropologist, 16 (January-March, 1914), pp. 1-25, is the early classic paper on the role of "physical" means of transportation as a facilitator of diffusion. Also see the later study of Erna Gunther, "The Westward Movement of Some Plains Traits," American Anthropologist, 52 (April-June, 1950), pp. 174-180. 40 See, for good examples, Stuart Rice, Quantitative Methods in Politics, New York: Knopf, 1928, pp. 154-155; A. T. and G. M. Culwick, "Culture Contact on the Fringe of Civilization," Africa, 8 (April, 1935), pp. 163-170; and, more recently, Charles J. Erasmus, --------------------- While a concern with channels is the predominant emphasis in several fields, it is a conspicuous lack in several others. Early anthropological studies, particularly those dealing with historical instances of diffusion, have been criticized for their (necessary, in part) lack of attention to process. Still, there were occasional studies pointing to probable means of transportation and communication such as Wissler's research on the horse in relation to the diffusion of Plains Indian culture traits" or the analyses of the role of roads and highways by various authors,4° and there was even a noteworthy study of the personalities and roles of key agents in the transmission of change.'" More recent anthropological studies of acculturation of technical assistance campaigns have given close attention to the character of the contacts between donor and recipient societies, a subject to which we shall return in the section on social structure below. An interesting new development in decision-making research has been the attempt, by several rural sociologists, to explore the psychological stages of the decision-making process and then to discover which media function most effectively within each stage.'" For example, for the initial "awareness" stage of receiving information, the mass media are obviously more efficient than inter personal relations, but the reverse is true for the stage of "acceptance." The importance of this work is that it makes even more salient one of the central themes of this decision-making tradition, which is that the channels are better viewed as complementary rather than competitive. In other words, it has become clear to many of those who have studied the role of the media in the making of decisions that different media are appropriate for different tasks and, consequently, that there is little worth to the gross question, which medium is more effective? These studies begin to be more interesting when they are carried out within a larger framework of structural and cultural factors. Ryan and Gross, for ex ample, used the decision-making approach to confirm the hypothesis which they found implicit in the logistic growth curve obtained for the spread of hybrid corn: that early adopters influenced the acceptance of the new seed by later adopters." ------------------------- "Agricultural Changes in Haiti: Patterns of Resistance and Acceptance," Human Organization, 2 (Winter, 1952), pp. 20-26. Some of these studies, it should be noted, are concerned rather more with channels of distribution than with channels of communication. 41. Paul Radin, op. cit., pp. 1-22. More recent examples include Richard N. Adams, "Personnel in Culture Change," Social Forces, 30 (December, 1951), pp. 185-189; Homer G. Barnett, "Personal Conflicts and Social Change," Social Forces, 20 (December, 1941), pp. 160-171; Wesley L. Bliss, "In the Wake of the Wheel," in Spicer (ed.), Human Problems in Technological Change, pp. 23-32; Henry F. Dobyns, "Experiment in Conservation," in Spicer, ibid., pp. 209-223; Allan R. Holmberg, "The Wells That Failed," in Spicer, ibid., pp. 113 123; Bertram Hutchinson, "Some Social Consequences of Nineteenth Century Missionary Activity Among the South African Bantis," Africa, 27 (April, 1957), pp. 160-175; I. Schap era, "Cultural Changes in Tribal Life," in Schapera (ed.), The Banta-Speaking Tribes of South Africa, London: Routledge and Sons, Ltd., 1937; Orner Stewart, Washo-Northern Paiute Peyotism: A Study in Acculturation, Berkeley: University of California Press, 1944; Fred Voget, "Individual Motivation in the Diffusion of the Wind River Shoshone Sundance to the Crow Indians," American Anthropologist, 50 (October-December, 1948), pp. 634-646; Fred Voget, "A Shoshone Innovator," American Anthropologist, 52 (January-March, 1950), pp. 52-63. 42. See James H. Copp, Maurice L. Sill and Emory J. Brown, "The Function of Information Sources in the Farm Practice Adoption Process," Rural Sociology, 23 (June, 1958), pp. 146-157; and Everett M. Rogers and George M. Beal, "The Importance of Personal Influence in the Adoption of Technical Changes," Social Forces, 36 (May, 1957), pp. 329-334. 43. Ryan and Gross, op. cit. ------------------- What should be clear by now, however, is that the place of many of these channel studies needs to be re-conceptualized. To the extent that they focus on interpersonal channels-that is, on the "relay" functions of interpersonal net works-they are concerned with social structure. And, if the sequence of events is taken into account whereby some persons are influenced by the mass media and others influenced by other persons, we have the beginnings of a diffusion study. Ideally, a diffusion study should classify individuals according to their place in a social structure-that is, according to their relationships with other people. What we need to know is when this kind of differential placement in the social structure is also related to differential access to, or acceptance of, influence stemming from outside the group regardless of whether the channel of influence is television or a troubadour or a traveling salesman. Then, we want to know whether differential placement in relationship to others has something to do with passing on, or reinforcing, information concerning the innovation. Thus studies of "who influences whom" fall into place both as structural studies and channel studies. Their content ranges from the role of a prestigeful person in introducing the sundance to the Crow Indians 44 to the influential role of women with large families in the realm of marketing." In short, what is needed is a wedding of studies of the channels of decision making and the social-structural approach to the study of diffusion so that in fluence and innovation can be traced as to how they make their way into a social structure from "outside" and as they diffuse through the networks of communication "inside." 6. SOCIAL STRUCTURE. From the point of view of diffusion research, then, the social structure functions in several different ways. First of all, it constitutes a set of boundaries within which items diffuse. Secondly, as has already been demonstrated, the social structure describes the major channels of person-to person communication through which diffusion flows. Additionally, social struc ture has to do with the distribution and differentiation of statuses and roles and the characteristic patterns of interaction among the occupants of varying positions. At least as far as diffusion is concerned, each of these functions may be seen to follow from the definition of social structure in terms of the frequency and the character of interpersonal contacts. ------------------- 44. See Fred Voget, "Individual Motivation . . . ," op. cit. 45. See Katz and Lazarsfeld, op. cit., Part III. ----------------- Consider boundaries, for example. Apart from making it possible to talk about the rate and extent of spread of an item within a system, boundaries are of interest to diffusion research because the frequency and character of social relations across a boundary differ from those within a boundary. Some studies have taken as problematic the determination of the effective boundaries within which diffusion takes place. For example, in his pioneering study of the diffusion of political influence, Stuart Rice discovered that state boundary lines acted as barriers to the diffusion of political influence except, interestingly, when residents of both sides of a state boundary shared a common marketing area." A number of studies have dealt with the boundaries which arise in connection with systems of social and ethnic stratification. Acceptance of an innovation by a lower social stratum, for example, may block acceptance by higher strata and, by the same token, upper-status groups-as Gillin has shown-may actually try to block the diffusion of symbolically meaningful items to groups of lesser status. The approach fits very well with classical sociological ideas about fashion changes in stratified societies. The same kind of thinking is characteristic of studies which have treated intergroup cleavages and rivalries within societies as boundaries to diffusion.° Several studies have inquired into the strategies of boundary-maintenance: Freed, for example, has analyzed the ways in which the traditional Amish and Eastern-European Jewish communities managed to constitute social structures limiting incursions of influence from the world outside. Finally, a number of anthropologists have confronted the problem of classifying the character of the social relations that exist across boundaries. Spicer, for example, tries to classify the variable relations between the Spanish conquerors and certain Indian tribes in terms of dimensions such as directed vs. non-directed, forced vs. permissive, hostile vs. friendly, and the like. From an analysis of these social interrelations, and the communications channels which they imply, have come various ideas about the kinds of items and changes which are likely to be associated with them. -------------------- 46. Stuart Rice, op. Cit. 47. See John Gillin, "Parallel Cultures and the Inhibition to Acculturation in a Guate malan Community," Social Forces, 24 (October, 1945), pp. 1-14; on this same theme, see the theoretical discussion by George Devereux and Edwin Loeb, "Antagonistic Acculturation," American Sociological Review, 8 (April, 1943), pp. 133-147. 48. See Georg Simmel, "Fashion," reprinted in American Journal of Sociology, 62 (May, 1957), pp. 541-558. 49. For example, Homer Barnett, "Applied Anthropology in 1860," Applied Anthropology, 1 (April-June, 1942), pp. 19-32. " Stanley A. Freed, "Suggested Type Societies in Acculturation Studies," American Anthropologist, 59 (February, 1957), pp. 55-68. Also see Joseph W. Eaton, "Controlled Ac culturation: A Survival Technique of the Hutterites," American Sociological Review, 17 (June, 1952), pp. 333-340; and Eric Wolf, "Aspects of Group Relations in a Complex Society: Mexico," American Anthropologist, 58 (December, 1956), pp. 1065-1078. 61. Thus, the combination of directed, permissive, friendly, intense and intimate contacts in the case of the Cahita led to the "fusion" of native and donor cultural elements while, in the case of the Athabascan, undirected, unforced, but hostile, intermittent and impersonal relations led to what Spicer calls "reorientation" or the adoption of a limited number of traits which, however, were extensively modified by the recipient culture. See Edward H. Spicer. "Spanish-Indian Acculturation in the Southwest," op. cit. For a related attempt to classify types of intergroup relations, see Edward P. Dozier, "Forced and Permissive Acculturation," op. cit. ----------------------- Curiously, more work has been done on the implications for diffusion of the structure of social relations across boundaries than within boundaries. Certainly very few studies have been done on the basic problem of comparing the ways in which different kinds of structural arrangements within a group condition the diffusion of a given item. There are some notable exceptions, however. Larsen and Hill, for example, studied the differential patterns of spread of a message in a working class and in a college community, and also in summer-camp com munities of varying degrees of stability. Lionberger studied variations in the flow of information as between residents in matched neighborhoods and "non neighborhoods," and Stuart Dodd found that variations in social relations resulting from differences in city size and population density affect the rate and extent of diffusion of airborne leaflets. Asking a different question about social structural relations, Albert argues, on the basis of a comparison of the rate of acceptance of Europein influences in Ruanda and Urundi that, under certain conditions, innovation will diffuse more rapidly in more centrally organized societies. Oscar Lewis has reported several cases of attempted assistance to under developed communities where the social structure of these communities played a key role in the fate of the project. Based on his restudy of the Mexican village of Teporthin, Lewis describes an effort to introduce a modern medical service which encountered resistance from those sectors of the village that would now be called the power structure of the community." In a second study Lewis de scribes the strategic significance of intro-community factions and cleavages for the eventual fate of innovations entering a village in India." To the extent that a society is more complex, networks of social relations become increasingly specialized. Thus, in the study of the diffusion of new drugs among doctors, networks of professional relations and networks of social relations were both found to carry influence, though at rather different rates and at rather different phases of the diffusion process. A related point is made by Edmondson to the effect that the uniform rate of spread which he finds in his study of rates of culture trait diffusion in the Neolithic may be a product of the essential similarity in the roles of all potential adopters; he speculates that the rise of specialists might change the picture substantially." More typical of current diffusion research is the use of social-structural factors to classify individuals rather than groups, both in terms of relative status and in terms of differential roles. A large number of rural studies take account of such factors as size-of-farm, age, education, membership in formal organizations and the like." While it is true that, in general, these variables are related to the acceptance of innovation in predictable ways, there are occasional surprises. -------------------- 52 See Otto N. Larsen and Richard J. Hill, "Mass Media and Interpersonal Communication," American Sociological Review, 19 (August, 1954), pp. 426-433; and "Social Structure and Interpersonal Communication," American Journal of Sociology, 63 (March, 1958), pp. 497-505. 53 Herbert F. Lionberger and Edward Hassinger, "Neighborhoods as a Factor in the Diffusion of Farm Information in a Northeast Missouri Farming Community," Rural Sociology, 19 (December, 1954), pp. 377-384. 54. See Stuart Dodd, "Formulas for Testing Opinions," Public Opinion Quarterly, 22 (Winter, 1958-59), pp. 537-554. 55. Ethel M. Albert, "Socio-Political Organization and Receptivity to Change: Some Dif ferences between Ruanda and Urundi," Southwestern Journal of Anthropology, 16 (Spring, 1960), pp. 46-74. 56. Oscar Lewis, "Medicine and Politics in a Mexican Village," in Benjamin Paul (ed.), Health, Culture and Community, op. cit., pp. 403-434. 57. Oscar Lewis, Group Dynamics in a North-Indian Village, A Study of Factions, New Delhi, India: Program Evaluation Organization, Planning Commission, 1954. The impor tance of social cleavages and factions in relation to the adoption and use of new items is sug gested by other studies as well, including A. R. Holmberg, "The Wells That Failed," in Edward Spicer (ed.), Human Problems in Technological Change, New York: Russell Sage Foundation, 1952, pp. 113-123; and also J. D. N. Versalius, "Social Factors in Asian Rural Development," Pacific Affairs, 30 (June, 1957), pp. 160-172. --------------------- A number of studies have shown that older people are more likely to accept certain innovations (those that contain a "revivalistic" element, for example) 61 and, similarly, another study found persons of lesser education to be earlier acceptors of the Salk Vaccine under certain circumstances." It is true that these standard variables do account for a considerable part of the variance in many studies, but they leave very many questions unanswered. And there are, of course, other structural variables which have been examined. Thus, Wilkening has studied the effect on innovation in farming of authoritarian vs. non-authori tarian family heads." Larsen and Hill, and Lionberger, are concerned with the ways in which social status within a primary group makes people differentially accessible to others both inside and outside the group." The study of the diffusion of a new drug among physicians focuses on the consequences of differential inte gration in the medical community for time-of-adoption," while a pioneering study in the field of marketing is concerned with the influence of a composite variable called "mobility" on time-of-adoption of new consumer goods. By the same token, group members have been studied in terms of the frequency and character of their contacts outside the group. Rural-sociological studies have taken accounts of such things as trips to the city, visiting outside the region, and personal contacts with agents of change such as salesmen, county agents and others who come into the community from the "outside world." Certain anthropologically oriented studies of technological change in developing areas have taken similar account of contacts outside the community as a factor making for individual differences in the acceptance of innovation." This kind of thinking, of course, leads directly to questions concerning the applicability of the hypothesis of the "two-step flow of communication" not only to mass com munications but to interpersonal diffusion as well: Does influence tend to flow from individuals with relatively more contact with the "outside world" (not only the mass media) to those who stay "at home?" ------------------ 58. Coleman, Katz and Menzel, op. cit. 59. Edmondson, op. cit. 60. For the influence of such variables on the acceptance of new farm practices see Lion berger, Adoption of New Ideas and Practices, op. cit., Chaps. 8 and 9. al See Fred Voget, "Individual Motivation . . . ," op. cit., and the literature on nativistic movements generally. in see John C. Belcher, "Acceptance of the Salk Polio Vaccine," Rural Sociology, 23 (June, 1958), pp. 158-170. Other studies of the diffusion of acceptance of the Salk Vaccine in other circumstances find the usual inverse relationship with education, social status, etc. Compare John A. Clausen, Morton A. Seidfeld and Leila C. Deasy, "Parent Attitudes toward Participation of Their Children in Polio Vaccine Trials," American Journal of Public Health, 44 (December, 1954), pp. 1526-1536. 63. Eugene A. Wilkening, "Changes in Farm Technology as Related to Familism, Family Decision Making and Family Integration," American Sociological Review, 19 (February, 1954), pp. 29-37. 64. Larsen and Hill, "Social Structure and Interpersonal Communication," op. cit.; Herbert F. Lionberger, "The Relation of Informal Social Groups to the Diffusion of Farm Information in a Northwest Missouri Farm Community," Rural Sociology, 19 (September, 1954), pp. 233-243. ----------------- 7. VALUE SYSTEMS. Social structures function, too, as anchorages for shared attitudes and values or, in other words, for culture. By the same token, roles are anchorages for certain individual differences in outlook and personality, though roles are not the only factor associated with personality. Attitudes, values and personality represent one of the major sets of variables that have been related to the acceptance of innovation and, if we consider them both at the level of the individual and of the group, it becomes possible to point out some interesting parallels between ostensibly unrelated traditions of research. ------------------ 65. Coleman, Katz and Menzel, op. cit. 66. See The Tastemakers, op. cit. 67. For example, Ryan and Gross, op. cit.: F. E. Emery and O. A. Oeser, "Information, Decision and Action, Melbourne: Melbourne University Press, 1958. 68. For example, Rose K. Goldsen and Max Rails, Factors Related to the Acceptance of Innovations in Bang Chan, Thailand, Ithaca: Cornell Thailand Project, Interim Reports Series, No. 3, 1957. 69. For a general discussion, and many specific examples, see Homer G. Barnett, Innova tion, New York: McGraw-Hill, 1953. This general conception has long been a fundamental postulate of anthropological thinking about cultural change. See, for example, in addition to the references cited in footnote 2, Boas' thinking in an early paper (1911) later reprinted in Race, Language and Culture, New York: Macmillan, 1940, p. 299. Somewhat later Linton ably stated the important elements of earlier anthropological thinking on this problem in The Study of Man, New York: Appleton Century Co., 1936.It should be noted that virtually from the beginning this conception of cultural compatibility has been applied to two distinctly different aspects of change phenomena. On the one hand, the compatibility conception has ... ----------------------- The central idea is that of "compatability" or "fit" between the culture of a group or the personality of the individual and the elements of a proposed innovation. On the group level, there are a number of anthropological studies under lining this principle." Among early studies, Lowie's, Wissler's, Radin's and Spier's studies on various aspects of diffusion among American Indians all emphasize the role of culture in making for selective borrowing. Somewhat later, Elsie Clews Parsons also stressed that traits were taken over by Mexican Indian townspeople from the Spanish and from others when they could "be fitted into an old form of behavior and (were) compatible with existing emotional attitudes." Since these early studies, anthropologically oriented research on diffusion has typically taken account of this principle. Furthermore, resistance to pro posed innovations as well as acceptance has often been explained in terms of this conception; in such cases, of course, the emphasis is upon the incompatibility between the receiving culture and the innovation. But all too few of these studies are comparative in the sense of setting out to demonstrate that a given item is acceptable to relatively comparable groups which, however, differ in values. One such example may be found in Oliver's study of the greater acceptability of new plant foods in a community many of whose rituals centered on the pig as compared with a community where taro, a plant, was a center of ritual and an important element in many institutional relations.73 Hawley reports on a similar comparative situation where Catholicism found greater acceptance among the patrilineally oriented Eastern Pueblo but was incompatible with the matrilineally oriented Western Pueblo. In much the same way, Saxon Graham seeks to explain the differential penetration of television ------------------- ... been applied to the problem of what might be called "symbolic" or "meaningful" fit between an innovation and the "mentality" of human targets of change; or perhaps more accurately, the compatibility between the meanings and symbolic significance of the innovation as perceived by the actors in question and their own system of values, attitudes and moods. On the other, the notion of compatibility has been applied to what might be referred to as "functional fit," i.e., the problem of the compatibility between the innovation and the adopting system viewed from the standpoint of the consequences of accepting and using the innovation. 70. Cf. footnote 2, above. 71. See Elsie Clews Parsons, Mida, Town of the Souls, Chicago: University of Chicago Press, 1936, p. 536. 72. For example, Charles J. Erasmus, op. cit., describes a situation among Haitian farmers, especially the backward ones, where the strong acceptance of a norm opposing "too much" material success acts to block and/or delay the diffusion of improved agricultural methods. Others who have emphasized this point include Charles P. Loomis and Glen Gresham, "The New Mexican Experiment in Village Rehabilitation," Applied Anthropology, 2 (June, 194.3), pp. 13-37; F. L. Bailey, "Suggested Techniques for Inducing Navaho Women to Accept Hospitalization During Childbirth," American Journal of Public Health, 38 (October, 1948), pp. 1418-1423; Morris E. Opler and Rudra Dott Singh, "Economic, Political and Social Change in a Village of North Central India." Human Organization, 11 (Summer, 1952), pp. 5-12; Bertram Hutchinson, "Some Social Consequences of Nineteenth Century Mis sionary Activity among the South African Bantu," Africa, 27 (April, 1957), pp. 160-175. 73. Douglas L. Oliver, "A Case of a Change in Food Habits in Bougainville, British Solomon Islands," Applied Anthropology, 1 (January-March, 1942), pp. 34-46. 74. Florence Hawley, "The Role of Pueblo Social Organization in the Dissemination of Catholicism," American Anthropologist, 48 (1946), pp. 407-415. ------------------------------- …and other leisure-time innovations in the middle and working classes in terms of the hospitality offered by the different sets of values of the two classes." On the individual level, the notion of compatibility, or fit, is equally applicable. Here can be located the whole tradition of motivation research in marketing. For motivation research is, in essence, the exploration of the symbolic meaning attributed by consumers to given items, seeking, ultimately, to tailor the item or its image to the consumer's personality." Studying the introduction of television in England, Himmelweit established that even when class member ship is held constant, different value orientations characterize early and late adopters. The former seemed more present-oriented while the latter were more future-oriented and perhaps inner-directed. In addition, rural sociologists have occasionally dealt with the problem of the functional compatibility of a new practice in relation to the personality characteristics of the individual." In any case, this classification brings very different research traditions into touch. Nevertheless, although the long-run aim may be the same, the dependent variables tend to be different. Hawley and Graham, for example, are concerned with the comparative extent of penetration of the item being studied in groups with different values. On the individual level, Himmelweit is concerned with the acceptance of TV by a given date. Motivation researchers, however, hardly ever study actual acceptance; their dependent variable is more likely to be "propensity to accept" and even that is often vaguely defined. Indeed, it may be said that this entire line of work requires that a distinction be made between the potential adopter's perception of the compatibility of an item and some objective evaluation of its compatibility, particularly over a longer period. This distinction parallels, to some extent, the earlier allusions to the difference between first use of an item and continued use. The item may be perceived as attractive to begin with, but experience with the item may involve unanticipated consequences which prove the longer-run incompatibility. Thus, the ease with which Puerto Rican women were willing to begin use of contraception does not jibe with the difficulties of of inducing them to continue regular use. In turn, this raises a more general… -------------------------- 75. Saxon Graham, "Class and Conservatism in the Adoption of Innovations," Human Relations, 9, 1 (1956), pp. 91-100. 76 See George H. Smiih, Motivation Research in Advertising and Marketing (New York: McGraw-Hill, 1953). An excellent example of work in this tradition is the early study of Maison Haire, "Projective Techniques in Marketing Research," Journal of Marketing, 14 (April, 1950), pp. 649-656, demonstrating that the initial resistance to instant coffee was based on an image that the product symbolized housewifely laziness. 77. Hilde Himmelweit, et al., Television and the Child, London: Oxford University Press, 1958. 78 For example, Irving A. Spaulding, "Farm Operator Time-Space Orientations and the Adoption of Recommended Farming Practices," Rhode Island Agricultural Experiment Sta tion Bulletin, No. 330, 1955; Everett M. Rogers, "Personality Correlates of the Adoption of Technical Practices," Rural Sociology, 22 (September, 1957), pp. 267-268. 79 See Hill, Stycos and Back, op. cit. Also see Apodaca, "Corn and Custom," in Spicer (ed.), op. cit. ---------------------------- …question concerning the tendency to overlook the fact that most innovative items consist of complex elements some of which may "fit" while others may not. Apart from the notion of functional fit, however, there are other subheadings within the cultural dimension which must be accounted for. Thus, there is a set of ideas, both on the group and on the individual level, which would seem to have more to do with a general orientation toward innovation than with the specific compatibility between certain innovations and certain values. Rural sociologists have conducted several studies of variations in ethnic attitudes toward innovation. On the individual level, too, early vs. late adopters, or adopters vs. non adopters, have been studied in terms of orientations such as sacred-secular, scientific-traditional, cosmopolitan-local and the like. CONCLUSIONS We have tried (1) to present an overview of the basic elements of the process of diffusion, and (2) to indicate, with respect to this accounting scheme, where each of a variety of research traditions has contributed as well as where it has fallen short, and (3) to specify problems which deserve further study. We have drawn specifically on the early work on diffusion in anthropology, sociology and education, and on more contemporary work stemming from the sociology of mass communication, rural sociology, studies of acculturation and of technical change, public health and marketing. We have hardly begun to explore the work in folklore, geography, archeology, and other fields. ----------------------- 80. For example, Harold A. Pederson, "Cultural Differences in the Acceptance of Recom mended Farm Practices," Rural Sociology, 16 (March, 1951), pp. 37-49; C. R. Hoffer, "Acceptance of Improved Farm Practices among Farmers of Dutch Descent," Michigan State College Agricultural Experiment Station, Special Bulletin No. 316, June, 1942. 81. For example, Emery and Oeser, op. cit.; Ryan and Gross, op. cit. 82 It deserves to be noted that, in 1952, a subcommittee of the Rural Sociological Society proposed a classification system for diffusion studies which resembles this one in part. It divided studies into those emphasizing (1) differential acceptance of farm practices as a function of status, role and motivation; (2) differential acceptance as a function of socio cultural systems; (3) diffusion as the study of cultural change; and (4) diffusion as a problem of the communication of information. The present paper differs, first of all, in that it advocates the integration of these several approaches in each study, though it also views the elements of the diffusion process as headings in terms of which to organize the various traditions of diffusion research. Secondly, as far as the specific classification schema is concerned, our inclination is to view categories one and two as parallel; accordingly we have grouped the individual (category one) and group (category two) factors together, dividing them only according to whether they are cultural or structural in emphasis. ---------------------- From the point of view of further development of the basic components, we have suggested (1) that the dependent variable, which we have been calling acceptance, must be more clearly defined; (2) that considerable ingenuity is needed to date the acceptance of innovations by their adopters, for time is the……key to diffusion research; (3) that considerable effort must be invested in the development of a "content analytic" scheme for classifying the item which is diffusing; (4) that attention must be given to the unit of adoption "required" by an item in the light of the unit which is "prescribed" or the unit which is the "target" of a communication campaign; (5) that interpersonal channels of communication must be viewed as elements of social structure; (6) that work is urgently needed on the comparative study of the same item diffusing in different social structures and, finally, (7) that the notion of "compatibility" between a given culture or personality and an item must be formulated much more strictly. From the point of view of the various traditions, we have tried to suggest how the work of each tradition contributes to a generic design for diffusion re search. Thus, anthropology brings into clear focus the group as the unit of adoption, and intergroup, rather than intra-group, contacts; it devotes considerable attention to the structure of social relations between donor and recipient as central to an understanding of the fate of an item moving from one group to the other; it raises the question of "levels" of acceptance. Another contribution of the work in anthropology centers around the concept of compatibility--that is, the extent to which a given culture is receptive to a given new item. But almost no attention is given to channels, and little information is available about the progress of an item over time. Early sociological work on diffusion also focused on corporate units of adoption (the municipality) as did educational research (the school system). In both these traditions, measures of time-of-adoption were explicitly formulated. Geo graphical proximity and urban-rural relations are the typical social structures in which channels of communication are thought, in some mysterious way, to inhere. More recent work in mass communication, rural sociology, public health and marketing has focused explicitly on the individual as the unit of adoption and on his perception of the channels of communication which influence his decision to adopt. Rural sociology has continually taken account of interpersonal relations as a channel but this has not been true of mass communication or marketing research until recently. In each of these fields, there appears to be a growing interest in exploring the social structures in which adopting units are linked, and to introduce time as a variable. If it becomes possible to combine this approach satisfactorily and still take account of the ways in which other channels of communication, including the mass media, impinge on these structures the problem of designing diffusion research for modern society will be well on its way to solution. But there are no easy answers so far. Also in Part 5:
|
Prev. | Next Top of Page | Index | Home |