Which measurement technique used to assess consumers evaluative criteria assumes consumers will not or Cannot state their evaluative criteria?

Citation:

John C. Philpot, Richard C. Reizenstein, and Daniel J. Sweeney (1972) ,"Identifying Determinants of Store Patronage Using Factor Analysis", in SV - Proceedings of the Third Annual Conference of the Association for Consumer Research, eds. M. Venkatesan, Chicago, IL : Association for Consumer Research, Pages: 201-212.

Proceedings of the Third Annual Conference of the Association for Consumer Research, 1972      Pages 201-212

IDENTIFYING DETERMINANTS OF STORE PATRONAGE USING FACTOR ANALYSIS

John C. Philpot

Richard C. Reizenstein

Daniel J. Sweeney

[John C. Philpot is Assistant Professor of Statistics. Richard C. Reizenstein and Daniel J. Sweeney are Assistant Professors of Marketing. All are of the University of Tennessee.]

INTRODUCTION

Although the selection of a preferred retail store is an important aspect of the consumer purchase decision process, it has received relatively little sophisticated attention in the relevant literature. Conceptually, the consumer's store patronage decision can be viewed as a comparison between certain evaluative criteria and certain perceived retail store characteristics (Engel, Kollat 8, Blackwell, 1968). The evaluative criteria represent the consumer's desires or expectations regarding various aspects of the retail store. These may be exemplified by store attribute importance variables such as importance of proximity to home, importance of layaway plans, and so forth. The comparison of the perceived characteristics of the retail store image with the consumer's evaluative criteria results in the identification of acceptable and unacceptable stores, indicating those stores which ultimately will and will not be patronized.

From a pragmatic point of view, the retailer needs to identify those evaluative criteria which consumers consider most important and which reflect dimensions of the store's image over which the retailer has some control. The retailer can then alter the characteristics of the store, attempting to make the perceived store image more consistent with the consumers' expectations (evaluative criteria).

Underlying the evaluative criteria are certain customer characteristics including shopping behavior patterns and personal demographic characteristics. These factors may influence the store patronage decision either directly in the form of established shopping habits, or indirectly by influencing the evaluative criteria. Thus, from both a conceptual and a pragmatic point of view, a key element of retail store patronage research is the identification of consistent sets of evaluative criteria used by different groups of customers to judge the acceptability of particular retail stores.

THE CURRENT STATE OF STORE PATRONAGE RESEARCH

Most of the store patronage research currently available can be classified either as store image studies or as retail segmentation studies. The store image studies tend to focus exclusively on the perceived image of specific retail stores and on the differences in perceived store images across consumers (Lazer & Wyckham, 1969; Martineau, 1958; Rich & Portis, 1964; Tillman, 1967). Little attention is given to the relative importance of the various dimensions of perceived store image to the consumer's store patronage decision. Moreover, there is typically no attempt to associate the store image dimensions with consistent sets of evaluative criteria.

Retail market segmentation studies all too often focus almost exclusively on the demographic and socioeconomic characteristics of retail store customers (Rich & Jain, 1968; Rich & Portis, 1963; Samli, 1970; Thompson, 1967). Rarely is any attempt made to identify retail market segments on the basis of consumers' expectations of retail stores. Exceptions include (Cravens & Cotham, 1970; Hughes, 1966; Sweeney & Reizenstein, 1972).

Some studies have been reported attempting to relate perceived store image dimensions to consumer demographic characteristics (Martineau, 1958; Rich 4 Jain, 1968; Rich X Portis, 1963 4 1964). Often, however, these analyses are based on cross tabulations of survey data with no indication of the relative strength of the observed relationships.

Furthermore, many patronage studies examine the consumer's choice between alternative stores with such widely differing appeals that the stores often cannot be considered directly competitive (Rich 8, Jain, }968; Rich 4 Portis, 1963). From the retailer's standpoint, the patronage decision in these cases is practically irrelevant.

Finally, retail patronage research has made only sparing use of multivariate analytical tools. Exceptions to this case include (Cravens & Cotham, 1970; Enis & Paul, 1970; Farley, 1968; Hughes, 1966; Sweeney & Reizenstein, 1972). The reliance on analyses that allow only one or two dimensions is a limiting condition on much of the store patronage research noted above.

Focus and Objectives of the Present Study

The present study focused on the store patronage decision of women's apparel purchasers among several directly competing women's specialty and department stores in a medium-sized Southeastern city. The objective of the present research was to identify the major determinants of women's apparel purchasers' store patronage decisions. More specifically, the study was designed to identify systematic relationships among consumers' store attribute expectations (evaluative criteria), their shopping behavior, their demographic characteristics and their preferences for particular women's apparel stores.

Methodology

A total of 500 questionnaires were mailed to a stratified random sample of households in the local retail trading area. One hundred forty-three useable questionnaires were returned, providing data on forty-six individual variables. These variables included:

1.  Evaluative Criteria: Relative importance of various store attributes, such as price, personnel, and services, to respondents' store preference and patronage decisions.

2.  Customer Characteristics: Respondents' shopping habits and patterns such as frequency of shopping trips and number of stores visited on each trip; and selected demographic characteristics.

3.   Respondents' preferences for and allocation of expenditures to specific womens' apparel and department stores.

The store attribute importance variables were rated by respondents using a seven point non-forced choice scale, anchored on a very unimportant to very important basis. All other variables were rated by checking the one of several discrete intervals which most closely reflected that respondent's shopping behavior, demographic characteristics, or store preference. (See further information in the Appendix)

Since the objective of the study was to search for meaningful ant stable relationships among the variables, the data were analyzed by means of factor analysis. Using squared multiple correlation coefficients as estimates of communality, principal factors were extracted. Those factor with eigenvalues greater than 1.0 were rotated using the varimax criterion, and a twelve factor model was developed. This model summarized 56.0% of the total variation among the forty-six variables and proved to be highly consistent with a principal-component solution.

ANALYSIS AND INTERPRETATION

Factor analysis is a multi-faceted technique. It enables us to express one set of variables in terms of a smaller set of factor variables, and is an approach commonly used for exploratory purposes as well as for tests of hypotheses. Farley, for example, has employed factor analysis in an exploratory manner to identify the determinants of supermarket patronage decision (Farley, 1968). Factor analysis has been employed in this study in a similar context, that is, simply as an exploratory vehicle to determine what types of stable store patronage factors exist.

The dimensions of a given factor can be identified in terms of the variables having heavy loadings on that factor. In addition, the factor axes may, in some cases, represent the clustering of related variables. For example, a set of correlated variables which is not correlated with the remainder of the variables in a study will constitute a separate factor (or subset of factors) in an orthogonal factor analysis of appropriate dimensions. While real data does not meet this condition, a factor may consist of a small set of moderately correlated variables whose members are weakly correlated with most other variables in the study. Such factors have more stability than those consisting of variables that are not directly correlated.

The interpretations that follow make use of these observations. The factor loadings have been examined in conjunction with the original correlation matrix, because correlations between particular pairs of variables loading on a factor are not necessary and cannot be assumed.

Of the twelve factors derived, eight will be examined closely. The remaining four factors, Factors 9, 10, 11, and 12, do not warrant comprehensive treatment as they cannot be clearly interpreted in terms of their component variables.

Factor 1 contains heavy loadings on variables related to age: age of major purchaser of women's apparel, males over 35, females over 35, males 18-35, and females 18-35. Other consumer demographics include family income and length of time as area resident. Shopping behavior variables of distance of store from home and length of time patronizing favorite store are of some importance. Availability of layaway plans is the only store attribute importance variable with a loading greater than .25. Preference for and expenditures in Store A are highly correlated with the dimension of this factor.

Interpretation of this factor suggests that the older purchaser of women's apparel, a loyal patron of her favorite store, a member of an older family, established in the community tends to avoid Store A. This consumer does not find availability of layaway plans an attractive store attribute.

TABLE 1

FACTOR LOADINGS FOR SELECTED VARIABLES - FACTOR 1

Opposed to this shopper is the younger (18-35 years of age), women's apparel purchaser, newly arrived, less affluent, less store loyal, more oriented toward availability of layaway plans. This latter consumer tends to prefer Store A, a contention supported by the fact that this store does, in fact, deal in sporty clothing with its primary appeal directed at the younger shopper.

Factor 2 comprises heavY loadings primarily on store attribute importance variables identified with a store's proximity to a consumer's home, to other women's apparel stores, and to other retail stores. The consumer demographic variable of female family members under 18 and preference for and expenditure in Store H are both related to this factor.

TABLE 2

FACTOR LOADINGS FOR SELECTED VARIABLES - FACTOR 2

It is apparent from this factor that shopping convenience, measured in terms of proximity,is an important evaluative criterion for some women's apparel shoppers. One store in particular, Store H, seems to be distinguishable from the others in terms of its failing to satisfy this criterion, as well as in its lack of appeal to younger women. As in the case of Factor 1, this interpretation is generally borne out by reality. Store H is located in a neighborhood shopping center with extremely poor traffic flow and parking. It is relatively isolated in regard to other such centers, and is close to only one other women's apparel store. Its merchandise is mainly oriented toward the more mature woman.

TABLE 3

FACTOR LOADINGS FOR SELECTED VARIABLES - FACTOR 3

Factor 3 is composed exclusively of store preference and expenditure variables. Store E has substantial loadings in both areas. Expenditure allocation in stores outside the study is of moderate importance.

The only interpretation which can validly be made here is quite obvious from the above data: Those consumers who prefer Store E purchase a substantial percentage of their women's apparel there. They do not spend a high percentage of their women's apparel dollars in stores outside the study.

Factor 4 loadings are most common on those variables related to income and expenditure. Store attribute importance variables of prices relative to prices of other stores and availability of layaway plans are also important. Moderate loadings are also indicated for expenditure allocation in Stores B, F, and G. Heaviest loadings are on average monthly expenditure on women's apparel and family income. Distance of favorite store from home and length of time as area resident are of some importance.

This factor suggests that consumers with high income who tend to

spend a large average amount on women's apparel each month do not consider availability of layaway plans or prices relative to prices of other stores to be important evaluative criteria in their store patronage decision. These consumers tend to allocate the greatest percentage of their women's apparel purchases to Stores B, F, and G. Conversely, those consumers who do consider price relative to prices of other stores and availability of layaway plans as important store attributes would most likely have lower incomes, spend less per month on women's apparel, and would allocate a much lower percentage (if any) of their women's apparel expenditures to Stores B, F, and G. This interpretation is fully supported by the fact that, of the eight specific stores studied, the three noted above are considered the most "exclusive," and carry the highest priced merchandise.

TABLE 4

FACTOR LOADINGS FOR SELECTED VARIABLES - FACTOR 4

TABLE 5

FACTOR LOADINGS FOR SELECTED VARIABLES - FACTOR 5

Factor 5 seems to focus primarily on variables related to monthly shopping frequency, women's apparel expenditures, and education. Heavy loadings are found on the shopping behavior variables related to the number of times the favorite store is visited monthly, the number of purchases made monthly at the favorite store, and the average monthly expenditure on women's apparel. The consumer demographic variable of educational level of the major purchaser of women's apparel is also heavily loaded. Finally, the store attribute importance variable of availability of specific brands is of some relevance, though its loading is much lower than those previously discussed.

This factor seems to indicate that those who shop and make purchases frequently in their favorite store, consequently spending a high average amount of money per month on women's apparel, tend to have a relatively low educational level compared to the remainder of the sample. These individuals seem (though the relationship is weak) to find availability of specific brands important, providing a possible insight into a reason for such frequent patronage of the favorite store. It also appears that a subset of those who do not shop or purchase frequently at their favorite store, and who do not spend a high average monthly sum on women's apparel, are relatively well educated. This segment does not seem to be as brand conscious as their lesser educated counterparts.

TABLE 6

FACTOR LOADINGS FOR SELECTED VARIABLES - FACTOR 6

Factor 6 seems to be a measure of the appeal of Store B contrasted with distance of the favorite store to home. The heaviest loadings are those on preference for and expenditure allocation to Store B. The shopping behavior variable of distance of favorite store from home is also quite noteworthy, with the store attribute importance variable of availability of specific brands being of some, though weaker importance.

This factor appears to indicate that part of the unique appeal of Store B is that distance from the consumer's home to the store is short. This is supported by the fact that Store B is one of the most exclusive specialty shops, and is located within five minutes of one of the most exclusive residential areas of the city. It can also be noted that it appears (though the relationship is weaker) that patrons of Store B tend to prefer specific brands.

TABLE 7

FACTOR LOADINGS FOR SELECTED VARIABLES - FACTOR 7

Factor 7 is characterized by variables related to young families with children. Strong factor loadings may be observed for the following consumer demographic variables: (1) Children at home; (2) Males, under 18; (3) Females, under 18; and (4) Age of major purchaser of women's apparel (though the loading here is weaker than in the previous three variables). The store attribute importance variable of range of sizes available was also considered quite important, as was availability of charge accounts.

This factor certainly seems to revolve around younger families, with children under 18 at home. These consumers do not tend to place much emphasis on the range of sizes available as an evaluative criterion for women's apparel. They do, however, seem to indicate the importance of the availability of charge accounts.

TABLE 8

FACTOR LOADINGS FOR SELECTED VARIABLES - FACTOR 8

Factor 8 contains extremely heavy loadings on Store D in regard to both store preference and expenditure allocation variables. Shopping behavior variables of number of purchases per month at favorite store and distance of favorite store from home are of some importance, though they are considerably less relevant than the variables directly relating to Store D.

The loadings for this factor indicate the overwhelming predominance of Store D in both store preference and expenditure allocation to favorite store. In addition, Store D customers tend to make more purchases per month and to travel a greater distance from home than those who do not patronize Store D. This interpretation is strongly supported by the fact that Store D is, by far, the dominant department store in the city, thus indicating that more people would travel further to patronize it. Moreover, the much wider variety of general merchandise carried in the department store would suggest a higher purchase frequency in Store D than in a store specializing exclusively in women's apparel.

CONCLUSIONS

Several general conclusions can be drawn from the study regarding the determinants of retail store patronage decisions, and the structure and methodology of future store patronage research.

1. The study identifies several sets of variables which appear to be related to the store patronage decision. In some cases the important variables relate to store attribute importance measures, in other cases to shopping behavior characteristics or to consumer demographic characteristics. The store attribute importance measures are particularly vital since these form the basis for identifying the evaluative criteria used by the consumer in selecting a preferred retail store. Having isolated the relevant evaluative criteria, the retailer should be able to redesign certain aspects of his store image to become more consistent with the consumer's store patronage expectations.

2. It appears that significant progress can be made in defining the determinants of store patronage if the analyst focuses on the interrelationships among a variety of different types of consumer characteristic and store attribute variables, using analytical tools capable of reflecting the multiple dimensions of the store patronage decision process. The present study indicates the value of one such multivariate technique, factor analysis, in the study of retail store Patronage decisions.

3. The present study suggests that the determinants of the store patronage decision are somewhat unique to stable groups of consumers as well as to specific stores, even though the stores are intensely competitive. This suggests that a highly differentiated market segmentation strategy might be very effective in the retail women's apparel market.

4. Additional research should be directed at identifying a more complete and sensitive set of store attribute and consumer characteristic variables for store patronage research. This need is underscored by the fact that five of the twelve factors identified in this study had heavy loadings on unique stores but only scattered loadings of moderate value on store attribute importance, shopping behavior, and consumer demographic variables.

REFERENCES

Cravens, David W., 4 Cotham, James C. Identifying Market Segments Using Canonical Correlation Analysis. A paper presented at the American Marketing Association 1970 Fall Educators Conference, Boston, Massachusetts, September, 1970.

Engel, James F., Kollat, David T., 4 Blackwell, Roger D. Consumer Behavior.New York: Holt, Rinehart, and Winston, Inc., 1968.

Enis, Ben M., Q Paul, Gordon W. 'Store Loyalty' as a Basis for Market Segmentation. Journal of Retailing, 1970, 46, 42-56.

Farley, John U. Dimensions of Supermarket Choice Patterns. Journal of Marketing Research, 1968, 5, 206-208.

Harman, Harry H. Modern Factor Analysis.Chicago: The University of Chicago Press, 1967.

Hughes, David G. Developing Marketing Strategy through Multiple Regression Journal of Marketing Research, 1966, 3, 412-415.

Lazer, William $ Wyckham, Robert G. Perceptual Segmentation of Department Store Markets. Journal of Retailing, 1969, 45, 3-14.

Martineau, Pierre. The Personality of the Retail Store. Harvard Business Review, 1958, 36, 47-55.

Rich, Stuart U. 4 Jain, Subhash C. Social Class and Life Cycle as Predictors of Shopping Behavior. Journal of Marketing Research, 1968, 5, 41-49.

Rich, Stuart U., X Portis, Bernard. Clues for Action from Shopper Preferences. Harvard Business Review, 1963, 41, 132-149.

Rich, Stuart U. 4 Portis, Bernard D. The Imageries of Department Stores. Journal of Marketing, 1964, 28, 10-15.

Samli, A. Coskun. Interrelationship between the Market Segments and the Buyer Behavior. Talk delivered before the X ESOMAR Seminar held at Lucerne,November 2-5, 1969, reprinted in David J. Rachman, Retail Management Strategy: Selected Readings. Englewood Cliffs: Prentice-Hall Inc., 1970, 97-112.

Sweeney, Daniel J. & Reizenstein, Richard C. Developing Retail Market Segmentation Strategy for a Women's Specialty Store Using Multiple Discriminant Analysis. A paper presented at the Fall Educator's Conference, The American Marketing Association, August, 1972.

Thompson, B. An Analysis of Supermarket Shopping Habits. Journal of Retailing, 1967, 43, pp. 17-29.

Tillman, Rollie. Semantic Differential Measurements of Consumer Images of Retail Stores. $he Southern Journal of Business, 1967, 2, 67-73.

APPENDIX

SCALES OF MEASUREMENT FOR VARIABLES

Store Attributes:

All variables (prices, brands, range of sizes, etc.) were on a 1-7 rank-order scale, where

1 = attribute very unimportant to the consumer

7 = attribute very important to the consumer

Shopping Behavior

All variables were on the positive integer scale, with integers corresponding to the following bases:

TABLE

*Unless otherwise specified, there was some compression of variable on the scale for the upper and/or lower values of the variable.

Consumer Demographics:

The following variables were on the positive integer scale, corresponding to the following bases:

TABLE

The remaining variables (Male over 35, Female over 35, Male 18-35, etc.) took on either 0 or 1 values, where:

0 = there was no member of the family corresponding to the age group

1 = there was at least one member of the family corresponding to the age groups

Store Preference:

All variables (Store A, Store B, etc.) indicated  consumer rankings of the respective stores on a -20 to -1 scale, where the largest value (-1) indicated the consumer's favorite store; the second largest value (-2) indicated the next most favored store, etc.  This scaling allowed the store preference variables to have, in general, the same signature for a given store as the corresponding expenditure allocation variables.

Expenditure Allocation:

All variable (Store A, Store B, etc.) consisted of the consumers' annual expenditure for women's apparel on a percentage scale, from 0 to 100.

----------------------------------------

Authors

John C. Philpot
Richard C. Reizenstein
Daniel J. Sweeney

Volume

SV - Proceedings of the Third Annual Conference of the Association for Consumer Research | 1972

Share Proceeding

Featured

“Eww, It Has a Face!” Anthropomorphizing Food Products Deteriorates Consumption Experience

Roland Schroll, University of Innsbruck, Austria

Read More

Featured

L5. Understanding the components and effects of the Omnichannel Seamless Experience.

PAULA RODRÍGUEZ-TORRICO, Universidad de Burgos (Spain)
Lauren Trabold, Manhattan College
Sonia San-Martín, University of Burgos (Spain)
Rebeca San José, University of Valladolid (Spain)

Read More

Featured

Approach and Loss Aversion: Consumer Responses to Approaching and Receding Stimuli in Advertising

Lana Mulier, Ghent University, Belgium
Iris Vermeir, Ghent University, Belgium
Hendrik Slabbinck, Ghent University, Belgium

Read More

Which measurement technique assess consumer evaluative criteria?

Which decision rule requires the consumer to rank the evaluative criteria in terms of their importance?

Which type of test is one in which the consumer is not aware of the products brand name?

What are the three types of consumer choice processes?