An Examination of Innovative Communicators, Opinion Leaders and Innovators For Men's Fashion Apparel

Elizabeth C. Hirschman, University of Pittsburgh
William O. Adcock, Georgia State University
ABSTRACT - Recent theoretical and empirical papers have suggested a need for more meaningful and useful adopter categories. The findings in this study provide support for the construct validity and pragmatic utility of an adopter category composed of individuals who are above average in both innovativeness and opinion leadership. A comparison of survey results from two socioeconomically diverse markets points to the potential danger of generalizing from research results obtained within a limited social environment.
[ to cite ]:
Elizabeth C. Hirschman and William O. Adcock (1978) ,"An Examination of Innovative Communicators, Opinion Leaders and Innovators For Men's Fashion Apparel", in NA - Advances in Consumer Research Volume 05, eds. Kent Hunt, Ann Abor, MI : Association for Consumer Research, Pages: 308-314.

Advances in Consumer Research Volume 5, 1978      Pages 308-314


Elizabeth C. Hirschman, University of Pittsburgh

William O. Adcock, Georgia State University


Recent theoretical and empirical papers have suggested a need for more meaningful and useful adopter categories. The findings in this study provide support for the construct validity and pragmatic utility of an adopter category composed of individuals who are above average in both innovativeness and opinion leadership. A comparison of survey results from two socioeconomically diverse markets points to the potential danger of generalizing from research results obtained within a limited social environment.


Diffusion of innovations is one of the most widely researched topics in the behavioral sciences. Rogers and Shoemaker (1971) compiled a bibliography of over 1,000 diffusion research studies which had been conducted in thirteen separate disciplines. Major diffusion research traditions exist in seven of these disciplines: anthropology, early sociology, rural sociology, education, medical sociology, communication, and marketing.

Despite the massive, combined research effort represented by the existing body of diffusion research, many observers are critical of the methodology, assumptions, and concepts associated with such studies, and many doubt that results are meaningful or useful. Rogers (1976) states that most diffusion research is lacking in originality and that contributions to more effective social programs have been few.

Taylor (1977) discusses the frustration felt by many consumer researchers that no meaningful set of relationships has yet been identified to characterize early triers of new products. Kotler and Zaltman (1976) note that research into early adopter characteristics has produced many contradictory results, and that few studies have proven useful in marketing planning. They also state that little information has been generated with regard to the sociological and psychological variables influencing the adoption decision.

Recently, Rogers (1976) has commented on the need for improved research designs, including network analysis and longitudinal research studies. He also stresses that scholars have accepted the diffusion paradigm so completely that they may be unable to recognize or to overcome possible conceptual biases.

The authors of this study concur with these scholars that diffusion research results have often been confusing and of limited usefulness, particularly in the marketing discipline. Improved research designs may lead to more meaningful results in the future. However, it appears that some of the basic concepts that have been widely accepted in diffusion research also need to be reexamined as a prelude to further analysis. This paper is intended to enlarge upon several recent theoretical and empirical efforts in determining relevant adopter categories.


Two categorizations are generally accepted as valid in diffusion theory. First, adoption patterns are assumed to follow a normal, bell-shaped distribution. Within this distribution five basic adopter categories are posited to exist: innovators (first 2.5 percent of population to adopt), early adopters (12.5 percent), early majority (34 percent), late majority (34 percent) and laggards (final 16 percent). The proportion of the population to be allocated to each adopter category reflects standard deviations within the normal distribution.

A number of marketing scholars have questioned the preciseness of this set of assumptions within a marketing context. Peterson (1973) states that a distribution in a marketing situation would probably be highly skewed, not normally distributed, because of the effects of increased numbers of producers and increased promotion as a new product began to win consumer acceptance. Peter-son also proposes the use of algorithms to identify categories in specific product classes in such a way as to minimize variance within groups. The use of such an algorithm could result in more or less than five categories in any given situation. Other marketing researchers have questioned the percentage distributions within categories. Robertson (1971) favors a ten percent allocation to the innovator category in marketing studies, while King (1963) classified 35 percent of his sample as "early buyers."

The second major categorization paradigm used in adoption theory concerns personal influence. A basic assumption of diffusion theory is that some individuals within a given social structure are influential in persuading others to adopt products. Attempts to identify the characteristics of these opinion leaders, however, have produced contradictory and inconclusive results (Zaltman and Stiff, 1973).


The overlap between these two sets of diffusion categories, innovativeness and opinion leadership, would intuitively appear to hold a great deal of potential for increasing understanding of the diffusion process. Yet surprisingly little empirical research has been specifically addressed to this issue. Robertson (1971) tabulated results from twenty-one marketing studies which had been conducted between 1954 and 1969, and reported that thirteen attempted to study relationships between innovativeness and opinion leadership. He determined that ten found a positive relationship, while three found no relationship. In no study was a negative relationship found.

Several recent studies of women's fashion adoption have reported disappointing results relative to correlation between innovativeness and opinion leadership. King (1963) reported that early and late buyers were not significantly different with respect to influence. Studies by Robertson and Myers (1969) and by Myers and Robertson (1972) found overlap between innovativeness and opinion leadership; but in each instance the overlap was described as moderate or small.

In a study of male fashion adoption Darden and Reynolds (1974) identified six adopter groups, as opposed to the five groups in classic diffusion theory. Discriminant analysis showed that opinion leadership was strongly associated with the most innovative adopter groups.

The authors were able to locate only two studies within the marketing discipline which were designed specifically to look at individuals who were above average in both innovativeness and opinion leadership. The first study was conducted by Summers (1971). He examined the overlap between innovativeness and opinion leadership found in generalized change agents for women's fashion apparel. He concluded that the relationship was not strong, although it was statistically significant. Baumgarten (1975) coined the term for these generalized change agents in a study he conducted on adoption processes for men's fashion apparel. Calling this adopter category, innovative communicators, Baumgarten concluded that they were differentiable from the general population and developed a psychographic profile of their characteristics.

The following year Midgley (1976) made an important theoretical contribution to the diffusion theory paradigm by formulating a three dimensional set of adopter categories. The three dimensions he proposes are: trial, acceptance/rejection and opinion leadership. His model incorporates four categories. Active adopters are individuals who have tried an innovation and will provide favorable information on it. Active rejectors are individuals who have tried an innovation, found it unsatisfactory, and will provide unfavorable information on it. Passives are those who have tried the innovation, but who will not provide information on it. The fourth group is termed potential adopters and includes those who have not yet tried the innovation.

Kotler and Zaltman (1976) added further sophistication to the issue of innovator-opinion leader overlap, when they formulated a theory of the best prospect. Noting that early adopters may not necessarily be the best target market segment for a new product, they provide a cost-benefit equation for determining the value of a prospect. The four factors that enter into this equation include (1) propensity to innovate, (2) amount the prospect is likely to purchase, (3) additional purchases that the prospect is likely to stimulate through interpersonal influence, and (4) the cost of communicating with the prospect.

The Kotler-Zaltman model does not require that the "best prospect" be high in both innovativeness and opinion leadership. Instead, the key is the nature of the interaction between the two characteristics, with purchase probabilities and communication costs also important to the economics of a target market strategy.

In summary, some empirical research has been conducted on the overlap between innovativeness and opinion leadership. Generally, these studies have identified relationships possessing statistical significance, but lacking in pragmatic utility. The theoretical work by Midgley and Kotler and Zaltman and the empirical work of Summers and Baumgarten, however, appear to provide a basis for defining a new, and pragmatically valuable, early adopter category - the innovative communicator.


The study reported here attempts to enlarge upon the earlier findings of Baumgarten (1975), with regard to male fashion "innovative communicators." These findings are certainly indicative of the possible existence and potential utility of the innovative communicator construct. However, some methodological limitations of Baumgarten's study prohibit the development of a generalized theory concerning the role of innovative communicators in the fashion apparel adoption process. Specifically, his data were limited to a sample of 389 unmarried male undergraduates at Purdue University. Thus, there is the potential for a restriction of range for the findings. Further, because data were gathered in only one setting, there can be no assurance as to their external validity. The necessity of assessing the external validity of findings from a study of this type is of central importance to the development of consumer behavior theory. As Shocker and Zaltman (1976) state:

In the marketing literature on consumer research, one still finds relatively little explicit concern with the internal and external validity attending the research itself...Issues of validation are very central to the development of theory and to the progression of research from mere ad hoc responses to specific inquiries toward the cohesive bodies of knowledge characteristic of disciplines.

A further limitation of Baumgarten's study was its failure to examine other possible adopter categories; all respondents who were not categorized as innovative communicators were combined into an "others" category. The study reported here uses adoption/influence patterns to specify four adopter categories. Innovative communicators, as defined in this study and in the Baumgarten study, are individuals who rank highly on both innovativeness and opinion leadership. A second group in this study, opinion leaders, includes respondents who score high on opinion leadership, but not on innovativeness. A third group, innovators includes individuals who score high on innovativeness, but not on opinion leadership. Finally, the general population includes those who do not score high on either construct.

The study reported here will also attempt to provide some preliminary examination of the four elements of the Kotler-Zaltman theory of the best prospect choice. In addition to measuring the interaction effect of overlapping innovativeness and opinion leadership, some data were gathered concerning media selection, store attribute evaluations and purchasing patterns.


Sampling Sites

Data for the research were gathered in two Southeastern cities during Spring, 1977. The cities and their respective sample sizes are given below:

Atlanta, Georgia     400

Augusta, Georgia   239

Total                       639

The cities and sample sizes were specified by the retail firm sponsoring the research. In Table 1 are given a set of statistics comparing the two cities on a variety of attributes.



Although the two cities are geographically adjacent, approximately 120 miles apart, they exhibit some large economic and demographic dissimilarities. The cities may be dichotomized using the statistics given in Table 1. At one end would be placed Atlanta, which is the larger, wealthier and more rapidly growing of the two cities. At the other end is Augusta with the smaller population, lower income per household and the slower rate of growth. This diversity between the cities comprising the sample should help provide a rich ground for assessing the external validity of the findings from this research.


To obtain the sample, a list of all telephone exchanges within each city's S.M.S.A. was compiled. By working closely with telephone company executives in each city, a recent (within the past two weeks) estimate was obtained of the number of telephone-equipped households within each exchange. The overall sample collected from each city was then apportioned according to the number of telephone households falling within a particular exchange.

A random digit dialing technique was used to gather the sample, which eliminated possible bias against new arrivals and unlisted numbers which can be caused by using the telephone directory. All interviewing was conducted from 6:00 PM to 9:30 PM weeknights to help insure an adequate representation of working persons. Trained interviewers employed by commercial field survey research firms were used to. conduct the interviews. In Table 2 are given the dates of interviewing for each city, the response rates, and the percentage of questionnaires which were validated.



One of the problems inherent in past diffusion research is that innovativeness and opinion leadership have been operationalized inconsistently. Rogers (1976), Kotler and Zaltman (1976), and Zaltman and Stiff (1973), among others, have criticized the measurement techniques used in most studies. These authors have advocated the use of network analysis to more accurately measure the constructs under study. The authors of this paper agree that such research designs are most desirable. As a practical matter, however, the resources needed for such studies may be unavailable for many researchers. As in the case of the study reported here, more limited measurement and operational schemes will continue to be used, but researchers should attempt to select designs that offer reasonable hope of construct, as well as face, validity.

The study reported in this paper utilized sets of self-designating questions to measure both fashion innovativeness and opinion leadership. The two sets of questions, three for each construct, are listed in Table 3. Responses to the question set for each construct were summed to create a composite score.



Male Fashion Innovative Communicators were operationally defined in this research as men who scored one standard deviation above the mean on both fashion innovativeness and fashion opinion leadership. These persons composed 6.8% of the sample (n = 16) in Augusta and 7.8% of the sample (n = 31) in Atlanta.

To describe the differentiating characteristics of the Male Fashion Innovative Communicator, consideration must be given to the groups against which it is appropriate to make comparisons. It was felt that for the purposes of assessing construct validity three relevant groups would be required. These three groups consisted of (1) Fashion Innovators - operationalized as persons scoring one standard deviation above the mean on fashion opinion leadership. (2) Fashion Opinion Leaders -operationalized as persons scoring one standard deviation above the mean on fashion opinion leadership but below one standard deviation above the mean on fashion innovativeness and (3) the General Population operationalized as persons scoring below one standard deviation above the mean on both fashion innovativeness and fashion opinion leadership.

The distributions obtained for the four groups are given in Table 4.




The variables used to compare male fashion innovative communicators with the three other groups of interest fell into five categories: (1) socioeconomic characteristics; (2) sociographic influences; (3) evaluations of retail store attributes; (4) media usage; and (5) store patronage.

The Atlanta sample received primary attention in the analysis reported here, rather than Augusta, because fashion interest was expected to be greater in the more cosmopolitan of the two cities. The analysis for each category of variables involved three major steps:

1. First, the data were examined to identify differences between the general population and the three construct groups (male fashion innovative communicators, male fashion opinion leaders, and male fashion innovators).

2. Next, the data were further examined to determine if differences existed among the three construct groups. Identification of significant differences would provide some indication of construct validity for the innovative communicator adopter category.

3. Finally, results from the Atlanta sample were compared with those in Augusta to test their external validity.

In each instance, the Chi Square statistic was used to test the significance of differences among groups (cut off at .10 level).


The variables examined included income, age, education, and length of residence in present city. Data reported in Table 5 shows that significant differences exist between the general population and the three construct groups only with respect to age. All three construct groups tend to include higher proportions of younger persons. This finding was also reflected in data from Augusta (Table 9). However, in neither Atlanta or Augusta were any significant differences found among the three construct groups with respect to age or any other socioeconomic variables.


Variables relating to participation in social activities and to memberships in groups were expected to provide valuable insights into the life styles and information transfer characteristics of male fashion adopters.

Social Activities

As shown in Table 6, all three construct groups are significantly different from the general population with regard to participation in a diverse group of activities. The construct groups are more likely than the general population to include individuals who entertain at home, attend movies, and attend concerts. Similar results were reflected by data from the Augusta sample.

The data in Table 6 also suggest a systematic and consistent pattern of differences among innovative communicators and other two construct groups. Although statistically significant differences exist for only four of the eight activities (attending concerts, gardening, golfing, and camping), the innovative communicator group contains the smallest proportion of non-participants for each of the activities reported here. This is true even though other groups contain higher proportions of very active participants for some activities. Thus, it appears that innovative communicators participate in a wide range of social activities, although levels of commitment to some activities may not be high. The Augusta data, however, do not reflect similar patterns of differences among the construct groups.

This finding may be attributed to three possible factors: (1) the small size of the Augusta sample introduces sampling error (2) the pattern of correlations found in the Atlanta data are spurious, or (3) the three constructs manifest themselves differently in various social settings. Atlanta is a large, wealthy and highly transient city relative to Augusta; it is possible that the great variance in social structure between these two cities may act to influence the form and function of social phenomena such as innovative communication, innovativeness and opinion leadership. Further, the types of and functions served by social activities and group memberships in the two cities may also be radically different.

If this third factor is indeed the most accurate rationale for the differences observed in the data, a strong argument could be made for greater attention being given to the impact of sociological variables on the adoption process, rather than limiting focus only to psychological correlates.

Group Membership. Respondents were asked to name groups and organizations to which they belong. The results indicate that unexpected behavior exists on the part of opinion leaders. Members of this group are significantly less likely than members of the other two construct groups to belong to social and recreational groups or organizations. Apparently, these individuals exercise personal influence within a less formalized or structured environment than might have been anticipated. Similar membership patterns are suggested by the Augusta data, but the results do not achieve statistical significance. Again, the impact of the social environment upon the manifestation of the construct is strongly suggested by the data.


Respondents were asked if each of ten store attributes was of much concern, moderate concern, or little concern to them in deciding where to shop. A comparison of evaluations of store attributes among the three construct groups and the general population is felt to have potential value for the development of more effective retail marketing strategies. One of the central criticisms of past diffusion research is that it has failed to demonstrate pragmatic utility (Kotler and Zaltman, 1976).

One finding in this area may be of value to both retail strategists and to consumer behavior theorists. As shown in Table 7, significant differences (.10 level) exist among the three construct groups for only one attribute, the store's guarantee, exchange and adjustment policies. The groups that exhibit innovative tendencies (innovative communicators and innovators) express more concern for guarantee, exchange and adjustment policies than do opinion leaders. Similar results, also significant at the .10 level, were found in Augusta.

This finding may indicate that an emphasis on guarantees or adjustment policies represents a form of risk reduction behavior on the part of male fashion innovators and innovative communicators. This would appear quite logical from the point of view of the fashion innovative consumer. Purchasing an unusually styled garment or one constructed from out-of-the-ordinary materials (for example, feather, fur or sequins) may activate both the social and economic components of perceived risk. Thus, if the store has a well-thought-of policy for merchandise return, exchange or repair, this may be an especially valuable attribute to the innovator.


Data were gathered only about newspaper subscriptions and radio station listenership, since these two media are of most pragmatic relevance from a retailing standpoint.

Data about newspaper readership included overall subscription rates and preferences for morning versus afternoon newspapers. In Atlanta, neither set of data identified differential tendencies toward newspaper readership among the four groups of interest in this study, (Table 8). However, in Augusta there were significant differences (.05) in subscription rates to the morning versus afternoon newspapers. These differences existed for both the three construct groups and the general population and among the three construct groups themselves. In Augusta the innovative communicators and opinion leaders were much more likely to subscribe to the afternoon newspaper than were innovators or the general population.

There are also significant and intuitively appealing differences among the four groups in terms of radio station preference. In both Atlanta and Augusta, the three construct groups display a distinct preference for stations featuring a progressive rock or jazz format, while the general population tends to be more partial to stations with classical or easy-listening type music. Differences among the four groups were significant at the .01 level in both Atlanta and Augusta. However, the Atlanta data do not reflect any significant differences among the three construct groups; while the Augusta data do.


It was expected that the four groups in question would display differential patronage patterns for the purchase of apparel. Respondents were asked where their last purchase had been made for men's suits and men's sportswear. Their responses were categorized into specialty and traditional department stores, national chain department stores, and discount stores.

Contrary to expectations, no significant differences among groups were found. It is possible that the "last purchase" specification was inadequate as an operationalization for patronage and that "usual store" would have provided somewhat different results. It is also possible that the store categories were inappropriate. In any event, survey data provide no basis for concluding that adopter groups prefer different types of stores.


The data discussed here identify significant patterns of differences at three levels of analysis: between the general population and construct groups; among construct groups; and between two socio-economically diverse markets.

Data show that Innovative Communicators, Opinion Leaders and Innovators tend to be younger than the general population, more likely to participate in several social activities, and prefer jazz/rock-oriented radio stations.

More importantly, significant differences were identified among the three construct groups. Innovative communicators are least likely to be non-participants in a wide range of activities. Opinion leaders are least likely to be members of social or recreational groups. Innovative communicators and innovators are more likely to express concern about guarantees and store adjustment policies than are opinion leaders. Findings of significant differences indicate that each group is unique and that the constructs possess more than face validity. It is also important to recognize that some findings in Atlanta were not replicated in Augusta. Differences in results may reflect the differences in the social structures and economic environments of the two cities. These differences suggest that it may be an error to attempt to generalize too widely from a single study pertaining to a limited social environment. Unless results are tested for external validity, generalizations are probably not justified.

This study provides empirical support for the Kotler-Zaltman theory of the best prospect choice. Since male fashion innovative communicators are above average in both opinion leadership and innovativeness and also express differential patterns of media preference, it should be possible to develop communications strategies directed toward these potential best prospects. However, further research will be needed to determine if shopping patterns and purchase propensities differentiate this group from other consumers.


The study reported here is intended to be transitional. The diffusion paradigm clearly needs to be examined from fresh perspectives and reformulated in a way that will be more useful to scholars in various disciplines as well as to marketing practitioners. Any new formulation of diffusion theory should involve a redefinition of adopter categories.

Findings that two unique groups of individuals, innovative communicators and innovators, exhibit innovative behavior help to explain the failure of earlier research to identify meaningful correlates of such behavior. A single set of predictor variables would be unlikely to describe members of both groups. For the same reason, findings that two unique groups of individuals, innovative communicators and opinion leaders, exercise personal influence explain the inconsistent results of studies intended to identify correlates for opinion leadership.

Findings from this study also suggest that failure to take sociological factors into account and failure to test for external validity have caused much of the previous research into innovation and opinion leadership to be of limited usefulness.

This study is not intended to be a definitive statement as to the nature of the adopter categories which will eventually prove to be most useful. The value of the findings reported here may simply be to suggest reasons why other studies have produced few useful insights and to demonstrate that new categories can be specified. Additional research involving network analysis and longitudinal designs may be needed to go beyond the two-dimensional scheme illustrated here, but such efforts are desirable if diffusion theory is to be made more meaningful.












Steven A. Baumgarten, "The Innovative Communicator in the Diffusion Process," Journal of Marketing Research, 12 (February 1975), 12-18.

William R. Darden and Fred D. Reynolds, "Backward Profiling of Male Innovators," Journal of Marketing Research, 11 (February 1974), 79-85.

Charles W. King, "Fashion Adoption: A Rebuttal to the 'Trickle Down' Theory," in Stephen A. Greyser, editor, Toward Scientific Marketing, Chicago: American Marketing Association, 1963, 108-125.

Philip Kotler and Gerald Zaltman, "Targeting Prospects for a New Product," Journal of Advertising Research, 16 (February 1976), 7-18.

David F. Midgley, "A Simple Mathematical Theory of Innovative Behavior," Journal of Consumer Research, 3 (June 1976), 31-41.

James H. Myers and Thomas S. Robertson, "Dimensions of Opinion Leadership," Journal of Marketing Research, 9 (February 1972) 41-46.

Robert A. Peterson, "A Note on Optimal Adopter Category Determination," Journal of Marketing Research, 10 (August 1973) 325-9.

Thomas S. Robertson, Innovative Behavior and Communication, New York: Holt, Rinehart and Winston, Inc., 1971.

Thomas S. Robertson and James H. Myers, "Personality Correlates of Opinion Leadership and Innovative Buying Behavior," Journal of Marketing Research, 6 (May 1969), 164-8.

Everett M. Rogers and Floyd Shoemaker, Communication of Innovations, New York: Free Press, 1971.

Everett M. Rogers, "New Product Adoption and Diffusion," Journal of Consumer Research, 2 (March 1976), 290-301.

Allen D. Shocker and Gerald Zaltman, "Validity Importance in Consumer Research: Some Pragmatic Issues," in William D. Perreault, Jr., editor, Advances in Consumer Research, Volume IV, Atlanta: Association for Consumer Research, 1977, 405-8.

John O. Summers, "Generalized Change Agents and Innovativeness," Journal of Marketing Research, 8 (August 1971), 313-6.

James W. Taylor, "A Striking Characteristic of Innovators,'' Journal of Marketing Research, 14 (February 1977), 104-107.

Gerald Zaltman and Ronald Stiff, "Theories of Diffusion,'' in Scott Ward and Thomas S. Robertson, editors, Consumer Behavior: Theoretical Sources, Englewood Cliffs, New Jersey: Prentice-Hall, Inc., 1973, 416-468.