Profiling the Male Fashion Innovator-- Another Step

John J. Painter, University of Utah
Kent L. Granzin, University of Utah
ABSTRACT - A cluster technique isolated three groups of adult males on the basis of 116 attributes of 10 clothing items recently acquired. The groups were compared in terms of their demographics, AIO's, and media preferences and habits. Results extended previous studies to provide an enlarged profile of the fashion innovator.
[ to cite ]:
John J. Painter and Kent L. Granzin (1976) ,"Profiling the Male Fashion Innovator-- Another Step", in NA - Advances in Consumer Research Volume 03, eds. Beverlee B. Anderson, Cincinnati, OH : Association for Consumer Research, Pages: 40-45.

Advances in Consumer Research Volume 3, 1976      Pages 40-45


John J. Painter, University of Utah

Kent L. Granzin, University of Utah


A cluster technique isolated three groups of adult males on the basis of 116 attributes of 10 clothing items recently acquired. The groups were compared in terms of their demographics, AIO's, and media preferences and habits. Results extended previous studies to provide an enlarged profile of the fashion innovator.


A number of recent studies have sought to identify and characterize the innovator (Baumgarten, 1975; Boone, 1970; Robertson, 1967; Robertson and Kennedy, 1968; Reynolds and Darden, 1972-73). While it would be desirable to generalize their findings beyond the product classes studied, it appears that in many cases such generalizations are unwarranted. Even within product classes noticeable differences appear to exist between men and women (Granzin and Painter, 1974; Painter and Pinegar, 1971).

This research concerns the important product category of men's clothing. Several studies of innovative behavior in the fashion area have generated profiles of the innovator that are enlightening, but incomplete. Granzin and Painter (1974) found demographic and media measures of the respondent useful for distinguishing innovators from non-innovators taken as a single group. Darden and Reynolds (1974) identified six consumer groups as based on innovative behavior. While they demonstrated the efficacy of demographic, AIO, and media variables for discriminating among the several groups, only one of the three product types used for group formation was clothing. Extension of their multiple group approach for the specific area of clothing purchases seems warranted.

The purpose of this study is to:

1. Provide further confirmation for the results of past studies in a multiple group context;

2. Consider new variables that may be useful in characterizing the male fashion innovator; and

3. Provide new, extended profiles of innovator and multiple non-innovator groups.

The results should provide marketers with insights that are useful in segmenting the market for men's clothing, whether their prime concern is the innovator or the non-innovator.


A random cluster sample of adult males in the Salt Lake City area was interviewed in their homes by student interviewers. City blocks were selected on an equal-probability basis from 1970 census tract listings, and a constant percentage of homes in each block contacted. In the event of turn-downs, interviewers selected the next dwelling on their previously constructed block maps. The survey resulted in 271 completed responses to the self-administered questionnaire.

The criterion for characterizing the innovativeness of a respondent was based on an extensive inventory of clothing characteristics. The inventory listed attributes of the last piece of clothing he purchased or requested in 10 different clothing categories. By checking the appropriate set of attributes, the respondent indicated the nature of a particular clothing item. For example, he characterized a suit coat with respect to: weave (woven or double knit); style (double- or single-breasted); lapel width (narrow, medium, or wide); cut (straight or shaped); back (belted or non-belted); and color (solid, print/pattern, or stripes). Other clothing categories were: sportcoats, pants for wear with suitcoat/sportcoat, casual slacks, dress shirts, casual shirts, sweaters, ties, belts, and shoes. Purchases and requests for purchase by someone else were used because men receive at least one-fourth of certain clothing items as gifts (Ryan, 1966).

Cluster analysis of the 271 respondents on the 116 different 0-1 scores was "seeded" with perceptions of the managers of 25 local retail clothing stores. These managers responded to the same list of attributes of clothing as did the respondents by indicating those attributes they deemed most characteristic of items that were innovative in terms of recency of market introduction. Perceptions of the 25 managers were introduced along with the responses of the 271 respondents into a cluster routine developed by Veldman (1967) and based on the Ward (1963) criterion for hierarchical grouping. The original 296 "groups" were systematically combined on the basis of the similarity of the clothing attributes they checked. By the time only nine groups remained, all 25 managers were classified into the same group, attesting to the homogeneity of their perceptions. Clustering proceeded until only three groups remained. One group contained the 25 managers and 33 respondents. The managers were removed from the group and the remaining individuals were dubbed innovators. A three-group cutoff was chosen because of the large increase in the Ward error term in going from three to two groups and because three groups appeared to furnish an interesting contrast with usual innovator/non-innovator two-group comparison common in the literature of innovative processes. The 33 innovators constituted 12.2 per cent of the sample, a proportion comparable to the 12 to 15 per cent Wasson (1971) suggests make up the class of early adopters for products in general.

Predictor variables were of three types: AIO's, demographics, and media preferences and habits. The 18 AIO's used in this research were selected to represent shopping behavior, recreational preferences, and occupational dress needs. The latter two categories were included because one's dress was presumed related to his activities. Demographics represent time-tested predictors of innovative behavior, although their predictive value has sometimes been of secondary importance. Eighteen demographic variables were included.

Media preferences and habits were included because of the necessity for marketing decision-makers to communicate with their target market(s). The instrument contained items relating to frequency and thoroughness of newspaper readership, number of magazines read regularly, and readership in different categories of magazines. Radio listening preferences referred to total hours a day of listening, preference for six classes of radio programming, and listening during eight periods distinguishable in a typical week. TV viewing related to total hours a day and preference for one of seven types of programming.

The three classes of predictor variables were analyzed separately by discriminant analyses on the three clothing groups. Use of the discriminant technique follows analyses of innovative behavior by Robertson and Kennedy (1968), Reynolds and Darden (1972-73), and Granzin and Painter (1974).


The cluster analysis generated three groups, which have been labeled A, B, and C. Table 1 represents the proportion of each group's members who purchased items possessing the listed attributes.

Group A (12.2 per cent of the sample) represents "innovators.'' They can be described as most likely to possess:

flare double-knit pants;

double-knit, wide lapel, shaped cut, single-breasted and solid color coats;

print or pattern, no stripes, double-knit shirts;

slip-on, non U-neck sweaters;

no suede, high-heel or two-tone shoes;

wide, no-stripe ties; and

some wide belts.

Group B, the largest of the three groups (45.7 per cent of the sample), may be characterized as clothing "conservatives'' who are most likely to own:

pants that are neither flare nor double-knit;

coats that are not double-knit, have medium lapels, are not shaped and single-breasted;

shirts that are conventional in style, not double-knit, not velour and without stripes;

no turtleneck sweaters;

no high-heel, suede or two-tone shoes; and

no wide belts.

Group C (42.1 per cent of the sample) is more of a challenge to describe. Although some members of this group do try some new clothing items, their wardrobes do not contain as broad a selection of innovative items as those of the Innovators. They appear to be more selective. These "middlers" may be characterized by their ownership of:

flare pants;

shaped, solid color, medium lapel coats;

no velour shirts;

some two-tone shoes;

wide ties; and

some wide belts.



Table 1 also indicates the results of the stepwise discriminant analysis of clothing attributes. The superscripts denote the order of entry of each attribute. Because the analysis operated on the same variables used in the group-forming cluster analysis, the presentation has no inferential value and is presented only to describe the clothing patterns characteristic of each group.


Table 2 presents means and univariate F ratios on the 18 demographic variables. Ten of the 18 variables were significant at the .05 level. The demographic variables are listed in Table 2 in the order that they entered the stepwise discriminant analysis. The functions were highly significant, as attested by a Wilks' lambda of .550 and a multivariate F of 4.85 (d.f. = 36/502; p = .000). Age, education, and stage of life cycle were the best discriminators. Income, though significant on a univariate basis, proved redundant to more significant items and entered relatively late in the stepwise process. A separate uni-variate analysis of annual clothing expenditures also indicated no significant difference among the three groups with respect to amounts spent on clothing.

Group A (Innovators) is low in marital status, number of children, age, and life cycle stage, although scarcely different from Group C in these four items. Highest in education, the Innovator group is intermediate in income, as might be expected from a younger group yet to reach its potential earning power. Intermediate in apartment renting, highest in house renting, this group is low in house ownership. The mobility suggested by this finding is entirely consistent with their younger age.

Group B (Conservatives) presents the reverse image of Group A. Highest in marital status, number of children, age, and life cycle stage--all by a wide margin--this group is simply older and more established in a life pattern than the other two groups. Lowest in education, this Conservative group has nonetheless achieved the highest income level. With a proportion of home ownership almost twice that of the next group, Group B can be portrayed as settled and having the least mobility.

Group C (Middlers) presents almost the same youthful pattern as Group A, but is lower on education level and income. Its higher proportion of apartment rental and lower inhabitation of a house, either through rental or ownership, suggests an even greater mobility than that seen for Group A. Group C apparently falls lower in the social order than Group A, and over time may be expected to converge upon the purchasing patterns of Group B to a greater extent.

Activities, Interests and Opinions

Table 3 presents means and univariate F ratios on 18 AIO's. Ten of the 18 variables were significant at the .05 level. Table 3 also indicates the results of the stepwise analysis of the AIO's, with each variable listed in the order of its entry. Again, the 18-variable discriminant functions were highly significant, with a Wilks' lambda of .635 and a multivariate F of 3.56 (d.f. = 36/502; p = .000). Purchase of new items before friends, enjoyment of active sports, and the purchase of name brand clothing to insure quality were the most important discriminators.

Group A (Innovators) indicates a relatively strong interest in trying new items on the market. This interest reflects more than curiosity, given the group's desire to look nice to women. Perhaps importantly, the group does not express a commensurately significant desire to dress well. One way of looking nice may be to buy name brand clothing. The Innovative group does not differ from the other groups in regard to the greater importance of quality and/or service over price. This group is least prone to dislike shopping for clothes, although it is not the most likely to know what it wants when going shopping. Perhaps clothes shopping is a pleasant experience for Group A. More interested in socializing, the group enjoys outdoor sports, quite possibly of a competitive nature. Because these recreation activities are relatively active, the group may generate more perspiration than accumulate actual dirt.



Group B (Conservatives) in most cases lies at the opposite end of the continua from Group A. Showing a relative disinterest in new items and in looking nice to women, this group lies intermediate on buying name brand clothing. The group is most likely to know what it wants when shopping, but this finding may be a reflection of its avoidance of new and different items. It is also highest on disliking the clothes shopping process. Least sociable, Group B is least likely to favor active and outdoor sports.

Group C (Middlers) presents, in most cases, a compromise between the two other groups. Its interest in new items, however, is high. Looking nice to women is a concern, but one not reflected in purchase of name brand clothing. Intermediate on dislike for clothes shopping, Group C is least likely to know what it wants when shopping. Almost as sociable as Group A, Group C is highest on active sports, relatively high on preference for outdoor recreation, and most likely to get dirty during these activities.

Media Preferences and Habits

Table 4 contains the means, proportions, and univariate F ratios for media habits and preferences. Of the 39 media variables, only nine had univariate F ratios significant at the .05 level. The 39-variable discriminant analysis gave a Wilks' lambda of .566 and a multivariate F of 1.939 (d.f. = 78/460; p = .000). Table 4 also presents the results of the stepwise analysis. Only thirteen of the variables entered here. Of the first five variables entering, two were related to radio (rock music and contemporary popular music), two were magazines (Reader's Digest and Playboy), and the other was the average amount of time spent watching television.

Time patterns of radio listening differ surprisingly little among the three groups, with the only exception being the 7 p.m. weekend evening time slot. However, Group A (Innovators) is high on listening to stations that play rock music and low on those featuring traditional (i.e., generally restrained) popular music. Lowest on interest in watching television, the Innovators show the highest desire to watch sports programming, which usually features competitive athletic contests. Most thorough in newspaper readership, Group A also leads in reading magazines featuring spectator sports and athletic events. The latter finding furnishes confirmation of its television preferences.

Group B (Conservatives) stands lowest in weekend evening radio listening, while its general listening pattern shows a clear preference for traditional popular music stations and an avoidance of rock stations. The group is highest in TV viewing, intermediate in sports interest, high in thoroughness of reading the newspaper, and lowest in readership of magazines featuring spectator sports and athletic events. Perhaps its interest in TV sports viewing is more compatible with a more sedentary life style than an active interest in competitive sports. The group's strong interest in Reader's Digest and clearly lowest readership of Playboy are consistent with the older and more conservative pattern emerging from this analysis.



Group C (Middlers) is highest on weekend-evening radio listening and shows the highest preference for rock music. Low in TV viewing, the group is also lowest in watching sports broadcasts. Least thorough in reading newspapers, Group C lags in Reader's Digest and magazines devoted to competitive sports. However, the group stands highest in Playboy readership.


The results of this study reveal profiles of the clothing innovator and non-innovator that have a good deal of similarity--and some dissimilarity--with the profiles generated by previous studies.



In general, this three-group analysis of the clothing market seems to present two definite extremes, as portrayed by Groups A and B. This study provides additional evidence that the innovator is most likely to be young, have few children, and be mobile. At the other extreme, it portrays the clothing conservative as least prone to innovation in general, much older, having a larger family, and not likely to be mobile. Doubt remains regarding the education level of the innovator because this study suggests the innovator has a higher education level than the non-innovator. On the other hand, the Darden and Reynolds study (1974) found the innovator to have a lower education level.

It also appears that clothing innovation is not importantly influenced by income level or clothing expenditures; rather, the issue is the way income is allocated to particular clothing purchases. This result is consistent with the finding of Darden and Reynolds (1974).

The profiles of the innovator and non-innovator are also expanded by use of AIO's. They characterize the innovator as more interested in brand name clothing and quality than in price. His desire to socialize is consistent with the results of a recent study of the "innovative communicator" by Baumgarten (1975). The innovator's strong predisposition to buy and try new items before friends and neighbors suggests that his innovative tendencies may extend beyond the clothing product category.

Differences in media habits and preferences go beyond magazine readership. For radio, the programming itself seems to be relevant. In particular, the rock music format seems to be an important variable, as it was in the Baumgarten study (1975). While the innovator reads the newspaper more thoroughly, he has a tendency to spend less time viewing television. He is more interested in sports programming than the non-innovator, and this interest in TV sports is consistent with his high level of interest in active sports participation. In general, the magazine readership results are consistent with findings of other studies (Darden and Reynolds, 1974; Granzin and Painter, 1974).

Given the results of this study, the innovator is fairly well profiled with respect to the simpler types of characterizing variables. In their future work researchers would do well to investigate more detailed items such as the form of the consumer's clothing search process and the way in which he decides on stores, brands, and clothing items.

Also, the usefulness of separating non-innovators into more than one group is questionable, given the static nature of the criterion measures. What is needed is a means for developing purchase groupings with more information content for the marketer. A methodological shift to longitudinal study of the patterns of clothing purchase over time could bring fresh insights. By recording the post-introduction timing of consumers' purchases, the time pattern of their adoption cycles could be obtained, and these patterns aggregated to provide more meaningful groupings of consumers than those obtained by present approaches. Such classification of purchase groups by the nature of their adoption patterns would be more useful to marketers seeking to formulate strategies for various stages of a clothing item's product life cycle.


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

Louis E. Boone, "The Search For the Consumer Innovator." Journal of Business, 43(April, 1970),134-140.

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

Kent L. Granzin and John J. Painter, "Characteristics of Male Fashion Innovators: A Discriminant Analysis," Mississippi Valley Journal of Business and Economics, 10(Fall, 1974),59-66.

John J. Painter and Max L. Pinegar, "Post-high Teens and Fashion Innovation," Journal of Marketing Research, 8(August, 1971),368-69.

Fred D. Reynolds and William R. Darden, "An Analysis of Selected Factors Associated with the Adoption of New Products," Mississippi Valley Journal of Business and Economics, 8(Winter, 1972-73),31-42.

Thomas S. Robertson, "Determinants of Innovative Behavior,'' Proceedings, Winter Conference, American Marketing Association, 1967, 328-32.

Thomas S. Robertson and James N. Kennedy, "Prediction of Consumer Innovators: Application of Multiple Discriminant Analysis," Journal of Marketing Research, 5(February, 1968),64-69.

Mary S. Ryan, Clothing: A Study in Human Behavior (New York: Holt, Rinehart and Winston, 1966).

Donald J. Veldman, Fortran Programming for the Behavioral Sciences (New York: Holt, Rinehart and Winston, 1967).

Joe H. Ward, Jr., "Hierarchical Grouping to Maximize an Objective Function," Journal of the American Statistical Association, 58(March, 1963),236-44.

Chester R. Wasson, Product Management: Product Life Cycles and Competitive Marketing Strategy (St. Charles, Illinois: Challenge Books, 1971).