Trickle down theory in fashion
Douglas J. Tigert, University of Toronto
Lawrence J. Ring, University of Toronto
Charles W. King, University of Toronto
[ to cite ]:
Douglas J. Tigert, Lawrence J. Ring, and Charles W. King (1976) ,"Fashion Involvement and Buying Behavior: a Methodological Study", in NA - Advances in Consumer Research Volume 03, eds. Beverlee B. Anderson, Cincinnati, OH : Association for Consumer Research, Pages: 46-52.
[ direct url ]:
Advances in Consumer Research Volume 3, 1976 Pages 46-52
FASHION INVOLVEMENT AND BUYING BEHAVIOR: A METHODOLOGICAL STUDY
Douglas J. Tigert, University of Toronto
Lawrence J. Ring, University of Toronto
Charles W. King, University of Toronto
This paper reports on continuing exploratory development of a comprehensive measure of "fashion involvement". The research presented here is the logical extension of a ten year tradition in fashion segmentation research and the fourth in a series of papers reporting on a major fashion research program in Canada. The "index of fashion involvement" developed here is validated with an independent AI0 measurement methodology and is utilized to demonstrate that the highly fashion involved consumer is also the heavy clothing fashion buyer.
The process by which new clothing and apparel concepts, "style statements", and tastes continually cycle across the population has been the subject of popular commentary for centuries. Aesthetic expression in dress and adornment is the classic statement of fashion, and has been the focus of a variety of strong methodological research papers over the past fifteen years. [For example see King (1963), Summers (1970), Baumgarten (1975), Grindereng, (1967), and Reynolds and Darden (1972).]
Contemporary fashion research indicates that clothing consumers are distributed across a broad spectrum of general fashion involvement or fashion consciousness (Sproles & King, 1973). Conceptually, this research has been developed and operationalized within the theoretical and methodological framework of the adoption and diffusion of innovations research tradition. Within that conceptual framework, contemporary fashion research has been, for the most part, focused on two key market segments: the fashion innovators or early adopters of new clothing concepts or "style statements"; and the fashion opinion leaders or interpersonal communicators, the local or peer group legitimizers of new clothing concepts.
The research presented here continues and builds upon those earlier efforts. Specifically, the paper is three-dimensional in focus:
1. The research further probes the concept of general fashion involvement. A "fashion involvement" index is developed and applied to similar male and female populations.
2. The issue of "index validation" is examined through the utilization of a comparative "fashion involvement" measurement by means of fashion specific life-style/AIO factor analytic techniques.
3. The concept of fashion market segmentation based on degree "fashion involvement" is probed with specific attention to unit and dollar volume fashion consumption.
CONTEMPORARY FASHION THEORY
Historically, contemporary fashion research has focused on the fashion innovator and the fashion opinion leader as prime sales targets and key links to the volume fashion market. Sophisticated methodological studies have been painstakingly designed to define, isolate, and profile these segments in a structured "block-by-block" manner over the past fifteen years, beginning with the research of King (1963).
In each of these cases, a variety of research measures have been utilized to develop an index (of innovative-ness, opinion leadership, etc.) upon which the population is continuously distributed. For example, King probed the traditional fashion adoption process model known as the "trickle down" theory by conducting an exploratory study of fashion change agents in the fashion adoption process within the context of women's millinery. The crucial variable in that Boston survey was the inclusion of the time of adoption in the fashion season in the research design.
Later, Summers (1970), focused on identifying and profiling opinion leaders/interpersonal communicators in the fashion adoption process. In this research, a sophisticated opinion leadership indexing technique was developed and tested. The basic conclusion in this and earlier research was that innovators and opinion leaders played key roles in directing fashion adoption and represented discrete market segments within social strata.
Carrying this structure a step further, Baumgarten (1975) researched the intersection of the fashion innovator and the fashion opinion leader: the "innovative communicator''. Theoretically, the innovative communicator performs jointly and simultaneously in the early buyer and new style communication and influential roles.
At the same time, King, Summers, Baumgarten, and others had been monitoring "fashion awareness" across a number of populations over several years. In the United States and Canada, as reported by King, Tigert and Ring (1975) a comparative fashion awareness data base spanning thirteen years has now been established.
The research described briefly here, and other related efforts indicate that "fashion innovativeness", "fashion opinion leadership", "fashion awareness", etc., exist as related but distinct continuums among the consumer population. Sproles and King (1973) suggest that such "fashion continuums" can he defined across a wide range of fashion-related activities and behavioral dimensions. What seems clear is that the population is differentially involved in fashion along a number of dimensions.
The Concept of Fashion Involvement
Given a number of empirically defined, distinct fashion behavioral dimensions, what is the joint and simultaneous interaction effect of that intersection? Baumgarten (1975) explored this question in part, by researching the intersection of the fashion innovator and the fashion opinion leader, the innovative communicator. In that research, the focus was on two dimensions of fashion related activity and behavior. It logically follows that there are several other additional dimensions which should also be included in this kind of analysis.
The concept of "fashion involvement" is based essentially on three propositions:
1. The population is distributed along a broad continuum in terms of fashion behavioral activities;
2. The population is also distributed on a unidimensional continuum for each of these fashion behavioral activities;
3. For several specific fashion behavioral activities, these continuums have been and can be researched and identified for specific geographic submarkets.
Theoretically, an overall fashion involvement continuum can be defined based on the aggregate effect of a variety of important fashion behavioral activities. Sproles and King (1973) suggest, based on previous fashion segmentation research, that there are at least five important dimensions of the aggregate fashion involvement continuum.
1. Fashion innovativeness and time of purchase. The continuum which ranges from the early adopting and experimenting consumer to the late buying, conservative consumer;
2. Fashion interpersonal communication. A continuous dimension which describes the relative communicative and influential power of the consuming population at conveying fashion information;
3. Fashion interest. A continuum ranging relatively from the highly interested fashion consumer to the totally non-interested buyer;
4. Fashion knowledgeability. Consumers range from those who are relatively knowledgeable about fashions, styles and trends to those who have no insight into the fashion arena;
5. Fashion awareness, and reaction to changing fashion trends. A continuum ranging from the consumer who is very actively monitoring the style trends to the totally non-aware individual.
As the first step toward building an aggregate index of fashion involvement, these five selected dimensions were utilized for initial analysis purposes. For each of the selected fashion behavioral dimensions, there exists a history of empirical research and tested measurement technology. Based on that research, [These measures have been developed and tested over the course of four major consumer research programs supervised by C.W. King. For the most recent work on the individual measures, see King, Tigert and Ring (1975) where the questions were tested individually preparatory to their aggregations into an index.] five individual questions, each a field tested measure of one of the five basic fashion behavioral dimensions, were selected for the development of "fashion involvement index".
Those questions are:
1. Fashion innovativeness and time of purchase. In general, would you say you buy new men's clothing fashions earlier in the season, about the same time, or later in the season than most other men?
Earlier in the season than most other men ..............
About the same time as most other men ..................
Later in the season than most other men ................
2. Fashion interpersonal communication. Would you say you give very little information, an average amount of information, or a great deal of information about new men's clothing fashions to your friends?
I give very little information to my friends ............
I give an average amount of information to my friends...
I give a great deal of information to my friends ........
3. Fashion interest. In general, would you say you are less interested, about as interested, or more interested in men's clothing fashions than most other men?
Less interested than most other men .....................
About as interested as most other men ...................
More interested than most other men .....................
4. Fashion knowledgeability. Compared with most other men, are you less likely, about as likely, or more likely to be asked for advice about new men's clothing fashions?
Less likely to be asked than most other men .............
About as likely to be asked as most other men ...........
More likely to be asked than most other men .............
5. Fashion awareness, and reaction to changing fashion trends. Which one of the statements below best describes your reaction to changing fashions in men's clothes? (Even though there may be no statement listed which exactly describes how you feel, make the best choice you can from the answers listed).
I read the fashion news regularly and try to keep my wardrobe up-to-date with the fashion trends .............
I keep up-to-date on all the fashion changes although I don't always attempt to dress according to these changes .................................................
I check to see what is currently fashionable only when I need to buy some new clothes ..........................
I don't pay much attention to fashion trends unless a major change takes place ................................
I am not at all interested in fashion trends ............
Simple sum scores across these five questions were computed for each consumer, resulting in a thirteen point continuum upon which the population is distributed.
The fashion awareness question has a five point response scale, compared to a three point scale for the remaining four dimensions. Therefore, the fashion awareness question has a proportionally higher weighting in the overall sum score. Further development using factor scores, standardized scores, and additional fashion dimensions is in progress.
This analysis is based on data gathered through the Toronto Retail Fashion Market Segmentation Research Program. [For a more complete description of the Toronto Retail Fashion Market Segmentation Research Program, see King, C.W., and L.J. Ring, "Retail Fashion Segmentation Research: Development and Implementation", in Bent Stidsen, editor, Marketing in the 1970's and Beyond, Canadian Association of Administrative Sciences (Proceedings), Edmonton, Canada, 1975.] This program developed from a collaboration by King, Tigert and Ring in the fashion arena, beginning with a 1974 study which integrated fashion segmentation dimensions with retail patronage analysis. From that initial effort a series of research projects evolved into an integrated research program. The broad objective of this program is to gain a better understanding of the dynamics of the fashion retailing process through the integration of contemporary fashion segmentation, lifestyle/AIO, and retail image methodologies within the broader program. This paper focuses on only one dimension of that process.
The data base for this research was collected in March/ April 1975 as part of the Toronto Retail Fashion Market Segmentation Research Program. The sample was part of a specially recruited panel of husband and wife pairs. The data were gathered by mailed self-administered questionnaires from an area probability sample in Census Metropolitan Toronto, Canada. A total of approximately 1000 husband and wife pairs completed and returned the questionnaires.
The objective of this section is to graphically illustrate the aggregate male and female fashion involvement sample distributions as determined by the fashion involvement index.
For each respondent, a simple sum score was calculated over the five selected questions listed earlier. Using a scale running from 5 (low) to 17 (high), the females and males were independently measured and analyzed. The distributions over the approximately 1000 respondents for both the male and female samples are displayed in Figure 1. The female distribution is uni-modal, and the male distribution is slightly bimodal. However, significantly different mean scores were obtained for the two samples. [The two sample means were five average standard errors apart. The statistical test for the hypothesis of no difference was rejected at the .001 level.] The female sample obtained a mean score of 9.7, while the male sample yielded a mean score of 8.4.
Distributionally, two observations are noteworthy.
1. Both the male and female distributions are strongly skewed toward the non-fashion involved end of the continuum. This may indicate that, in general, Toronto consumers are not highly fashion involved.
2. Toronto females are more "involved" in fashion than are Toronto males, based on mean sample differences.
VALIDATION OF THE FASHION INVOLVEMENT MEASURE
For this research, an attempt was made to validate the concept of fashion involvement, and this fashion involvement index by measuring it in a second and entirely independent and unrelated manner to what has historically been the case in fashion segmentation research. Factor analytic techniques have been utilized across a variety of fashion specific lifestyle/AI0 questions to obtain a "fashion involvement factor".
1975 TORONTO MALE AND FEMALE FASHION STUDIES: SAMPLE DISTRIBUTION (n = 1,000) ON FASHION INVOLVEMENT INDEX (SUM SCORE)
Over the past several years, the authors have conducted numerous in-depth, focused group discussions with men and women about fashion in both the United States and Canada. Based on that probing, a number of hypotheses have been generated about fashion attitudes and activities and fashion buying behavior which carry strong implications for fashion retailers.
Those hypotheses have been stated in 24 different lifestyle measures. The measures ranged across social attitudes (non-fashion specific) about unionism, sexual behavior, religion, and deviant behavior. Additionally, the measures covered a wide range of fashion related dimensions, such as commitment to fashionability, fashion trends, and clothes shopping. The measures have been operationalized around a six point (strongly agree to strongly disagree) Likert Scale. The questions were included in the questionnaire utilized in data collection for this paper. In addition, many of these same questions have been used in previous fashion related research efforts.
The responses to the 24 lifestyle/AIO questions were submitted to a principal factor analysis and varimax rotation. The first and strongest factor in both the male and female data is employed here as another measure of fashion involvement/consciousness and has been so labeled. The rotated factor loadings are reported in Table I for both the male and the female data. A similar factor has been produced and reproduced in several previous studies in both the United States and Canada (Wells & Tigert, 1971;and Tigert, 1973). This factor has been both highly reliable and stable regardless of the number of mix of lifestyle questions analyzed in any particular study. Additionally, the factor has been utilized as a predictor of fashion buying behavior, fashion magazine readership and other fashion attitudes.
TORONTO, 1975: FASHION INVOLVEMENT/CONSCIOUSNESS FACTOR
The Fashion Involvement Index vs. The Fashion Involvement Factor
Distributionally, both the fashion involvement index and the fashion involvement factor obtained higher mean scores for females than males. For the factor, females obtained a sample mean which was 6.5 percent higher than the male sample mean score (rejecting the no difference hypothesis of the .001 level). For the "fashion involvement index", the female mean was 10 percent higher than the male mean. In addition, the shapes of both sets of distributions were highly similar.
As a measure of that similarity, correlation analysis was run between the index and factor sum scores. The two measures obtained a correlation coefficient of .633 in the male case, and .566 in the female example.
Therefore, the conclusion is that while the index and the factor do not measure exactly the same thing, they are similar enough to suggest that the concept of "fashion involvement" is in fact a valid one.
FASHION INVOLVEMENT AND BUYING BEHAVIOR
To date, contemporary fashion researchers have devoted their primary research efforts toward the identification and profiling of key market segments thought to be driving or influencing the fashion adoption process. Innovators and opinion leaders have been profiled by several fashion research teams across a wide variety of sociological, psychological, lifestyle, demographic, and other characteristics.
What has not been explored in great depth, however, is the actual buying and purchasing behavior of highly fashion involved consumers. In a paper presented before the American Marketing Association in August, 1975, the authors probed retail clothing fashion store shopping behavior across several dimensions of fashion involvement. A methodology was developed for the application of contemporary fashion segmentation technology to a specific fashion retailing context.
In that Toronto-based research, it was clear that significant differences existed in the levels of fashion involvement among the patrons of major Toronto men's wear chains. For example, in general the fashion specialty chains had a much more "fashion involved" customer mix than did the mass merchandiser set of stores.
In order to further probe consumer fashion purchasing behavior, this section of the paper examines the relationship between fashion involvement and unit and dollar volume fashion buying behavior.
Toronto Male Clothes Buying
As a part of the 1975 Toronto Male Fashion Study which was described in the data base section of this paper, respondents were asked to estimate the number of units they purchased over the past year across 11 fashion product/object categories. In addition, respondents estimated the average price they paid across those same 11 categories.
For this paper, the 11 categories were collapsed into three major classes, "big ticket items", "dressy accessories'', and "casual wear". Three particular fashion product/objects were selected, one from each class, as representative of Toronto male fashion buying in general.
The analysis here focuses on suits, dress shirts, and casual slacks (defined as "jeans, corduroy, denim"). For each of these three categories, cross classification analysis was performed across unit and dollar volume purchased against "fashion involvement" as measured by the index described earlier in this paper.
As an operating hypothesis, the expectation was that the more highly fashion involved consumers would be heavier clothing buyers in terms of both price (price per unit bought) and volume (number of units bought) than less fashion involved individuals. That hypothesis was, in fact, supported. Those results are depicted in Tables 2, 3, 4.
For purposes of the cross classification analyses, the fashion involvement distribution was arbitrarily segmented into "low, medium, and high" fashion involved groupings. Those groupings are reflected in Figure 1 which was presented earlier.
The cross classification analysis sequence graphically illustrates some rather dramatic differences in fashion purchasing behavior between the "high" fashion involved group and the "medium to low" groups.
For example, turning first to suits as a fashion product/object category (Table 2), the "high" fashion involved consumers are clearly the heavy buyers in terms of both numbers of suits and in terms of dollars per suit. Twenty percent of "high" fashion involved people said they bought four or more suits in the past year. This is four times the number in either the "medium" or "low" fashion involved groups. In terms of prices, 47 percent of "high" fashion involved men said they paid over 0 per suit on average, while the "medium" and "low" fashion involved groups had less than 20 percent purchasing suits over 0. For both unit and dollar analyses, the differences between the groups were statically significant based on a chisquare test at the .001 level.
FASHION INVOLVEMENT AND SUIT BUYING
13.7 percent of the "Low" group said they purchased six or more dress shirts during the same period. Price-wise the differences are similar, with well over fifty percent of "high" fashion involved people paying the high prices, over twice as many as the "medium" and "low" groups.
In the casual slacks analysis sequence, Table 4, the differences were not as dramatic as in the suit and dress shirt analyses. However, the "high" fashion involved group still bought more pairs of casual slacks, and paid more money per pair of casual slacks than did the members of the "medium" and "low" fashion involved groups. The smaller differences here may be attributed to the fact that casual slacks are generally not considered as much of a "fashion" apparel item as suits and dressy shirts.
FASHION INVOLVEMENT AND DRESS SHIRT BUYING
With regard to the dress shirt analysis, Table 3, again the "high" fashion involved consumers are the heavy buyers. 51.1 percent of the "high" fashion involved group said they bought six or more dress shirts over the year. In contrast, only 29.6 percent of the "medium" group and Overall, the cross classification analyses have revealed that "high" fashion involved consumers, are much heavier clothing fashion buyers across most, if not all, classes of men's wear, than their less fashion involved counterparts.
FASHION INVOLVEMENT AND CASUAL SLACKS BUYING
CONCLUSIONS AND IMPLICATIONS
The concept of fashion involvement as a summary or cumulative statement of at least five dimensions of fashion adoption-related behavior was defined and researched. An index of fashion involvement was developed and the construct was validated with an independent fashion specific lifestyle/AIO factor analytic technique.
Based on this index, there are major differences among consumers in terms of fashion involvement. The distribution of the male and female populations in the Toronto fashion market reflected a wide range of fashion involvement. Toronto males, however, were lower in fashion involvement than their female counterparts.
The high fashion involved consumer has been the historical interest of fashion researchers, as the drivers and influentials and legitimitizers of the overall fashion adoption process. This concept has been supported in this paper as a result of a strong and significant relationship identified between fashion involvement and unit and dollar clothing fashion buying behavior.
The findings obtained here strongly suggest that in addition to being the driving force in the fashion adoption process, the highly fashion involved consumers represent an important market as heavy clothing buyers. While, the highly fashion involved group is relatively small vis-a-vis the total population, that group is much larger in terms of proportionate share of clothing fashions purchased. For example, the "high" fashion involved group of suit buying population (4 percent) bought over 10 percent of the suits.
In summary, this paper is part of an on-going research program focusing on fashion adoption. More specifically, this paper has:
1. Developed a fashion involvement index which has face validity and is verified as a construct in independent lifestyle/AIO research.
2. The fashion involvement construct suggests that men and women are different in terms of their fashion involvement.
3. The "high" fashion involved consumer is important to monitor for the fashion industry and particularly the fashion merchant. The "high" fashion involved consumer is the fashion leader in innovativeness, early trial, and interpersonal communication of fashion information. This market segment is also a market target representing a disproportionately high buying segment of the population.
Further research in this program will explore the dynamics of the fashion buying process in greater detail.
Baumgarten, Steven A. "The Innovative Communicator in the Diffusion Process." Journal of Marketing Research, 12 (February 1975), 12-18.
Grindereng, Margaret P. "Fashion Diffusion." Journal of Home Economics, 59 (March 1967), 171-174.
King, Charles W. "Fashion Adoption: A Rebuttal to the 'Trickle Down' Theory." In Stephen A. Greyser, editor, Toward Scientific Marketing, Chicago. American Marketing Association, 1963, 108-125.
King, Charles W., and L.J. Ring "Retail Fashion Segmentation Research: Development and Implementation." In Bent Stidsen, editor, Marketing in the 1970's and Beyond, Edmonton, Canada. Canadian Association of Administrative Sciences, 1975.
King, Charles W., Tigert, Douglas J. and Ring, L.J., "Contemporary Fashion Theory and Retail Shopping Behavior: A Segmentation Analysis." In the 1975 Combined Proceedings, American Marketing Association.
Reynolds, Fred D., and Williams R. Darden. "Why the Midi Failed." Journal of Advertising Research, 12 (August 1972), 39-44.
Sproles, George B., and Charles W. King. "The Consumer Fashion Change Agent: A Theoretical Conceptualization and Empirical Identification." Institute for Research in the Behavioral, Economic, and Management Sciences, Purdue University, Paper No. 433, (November 1973).
Summers, John O. "The Identity of Women's Clothing Fashion Opinion Leaders." Journal of Marketing Research, 7 (May 1970), 178-185.
Tigert, Douglas J. "Can a Separate Marketing Strategy for French Canada be Justified: Profiling English-French Markets Through Life Style Analysis." In Canadian Marketing: Problems and Prospects, editor Thompson and Leighton, Toronto, Canada. Wiley Publishers of Canada, Ltd., 1973.
Wells, William D., and Douglas J. Tigert. "Activities, Interests, and Opinions." Journal of Advertising Research, 11 (August 1971), 27-35.