Shopping Area Image: Its Factor Analytic Structure and Relationships With Shopping Trips and Expenditure Behavior

ABSTRACT - Few studies have focused on the image of a shopping area and its relationship with shopping trips and expenditures. This research focuses on the development of a scale to measure -the image of a downtown shopping area. Of the four component scales identified, two - the operational and facilitative factors - were found to significantly affect shopping trips and expenditures


Chow Hou Wee (1986) ,"Shopping Area Image: Its Factor Analytic Structure and Relationships With Shopping Trips and Expenditure Behavior", in NA - Advances in Consumer Research Volume 13, eds. Richard J. Lutz, Provo, UT : Association for Consumer Research, Pages: 48-52.

Advances in Consumer Research Volume 13, 1986      Pages 48-52


Chow Hou Wee, National University of Singapore


Few studies have focused on the image of a shopping area and its relationship with shopping trips and expenditures. This research focuses on the development of a scale to measure -the image of a downtown shopping area. Of the four component scales identified, two - the operational and facilitative factors - were found to significantly affect shopping trips and expenditures


Marketing research on image had historically been confined to the store level. However, with the emergence of more and more shopping centers as business entities, there is a need to focus research on the image of a shopping area. In fact, there is no theoretical reason why image research at the store level cannot be extended and applied to the shopping area level. The importance and relevance of image research to shopping area patronage was recognized some 15 years ago by Moore and Mason (1969):

"Socio-economic variables do not satisfactorily explain the retail centre patronage decisions of the residents of the study area. It may be inferred that psychological or attitudinal differences of the residents are perhaps of more importance.

With the development of shopping centers, studies have begun to focus on image-like variables of shopping areas (Frederick et al. 1975; Carter 1978, 1981; Hauser and Koppelman 1979; O'Neill and Hawkins 1980; Berman 1983). Studies by Bellenger, Robertson and Greenberg (1977), and Gentry and Burns (1977) have confirmed the importance of image-like variables in shopping center patronage.

However, the most significant study on shopping area image thus far has been that of Houston and Nevin (1980). In their study of the downtown area and four shopping centres on 16 image items, Houston and Nevin used factor analysis to identify three major dimensions or factors of shopping area image. The first factor consisted of six items -- quality of stores, variety of stores, merchandise quality, product selection, special sales/promotion, and great place to spend a few hours -- which were related to the assortment of benefits offered by the area. The second factor (6 items) consisted of features that helped to ease the shopping effort -- parking facilities, availability of lunch/refreshment, comfort areas, easy to take children, layout of area, and special events/exhibits - and vas named the facilitative nature of the area. The third factor (4 items) -- general price level, atmosphere, store personnel, and conservative -- were associated with positioning of the area as an integrated complex of stores, and was named market posture.


There are several areas of concern that still need to be addressed with regard to the study of shopping area image. The first concern involves the appropriateness of applying the ideas drawn from the store image literature to that of shopping area research. Houston and Nevin (1980) developed their scale from a review of earlier store image research and discussion with shopping center managers. Whether the items were appropriate, especially in the eyes of the consumers, were not considered. This limitation vas clearly recognized by Howell and Rogers (1980). Thus, to begin with, it may be more appropriate to sample a wider domain of items, especially at such an exploratory level, in order to have a better understanding of the underlying dimensions of shopping area image.

The second concern involves the issue of familiarity. Acito and Anderson (1979) found that image was more differentiated, better articulated and of higher dimensionality for recent shoppers compared with non recent shoppers of a retail store. The same concern was shared by Hirschman (1981). In essence, image research that is based on ratings by consumers without taking into account the extent of their familiarity with the stores or areas is not likely to provide meaningful results that managers can act upon. For example, can one rely on the ratings of consumers, whether favorable or unfavorable, if they have never been to particular shopping areas, or only been there once or twice a year? When one considers that respondents are typically asked to react to some 15 to 20 items on the image scale, the problem is further compounded. It can be argued that the image of a shopping area may be more complex than that of a single store, since a shopping area is a conglomerate of different kinds of stores that offer a wide variety of products and services. In fact, in a recent study, Wee (1985) found that the factor analytic structure of the image scale for a shopping area differed with regard to recency effects and the size of the shopping area. Thus, the need for the consumer to be familiar with the shopping area is crucial to any assessment of its image.

Finally, the relationship between shopping area image and patronage behaviour, especially in terms of actual consumption behavior like number of trips made and the amount of expenditure spent at the area, needs to be better established. The findings by Nevin and Houston (1980) should not be taken as conclusive in that they included a variable on preferred store, measured in a dichotomous basis, in their study. To the extent that the image of a shopping center is correlated with the image of that store (whether or not shopping center image actually exists), the significance of shopping center image will lessen. Thus, it is possible for a shopping center image to exist and help determine shopping behavior, yet emerge as insignificant from Nevin and Houston's analysis.


The purpose of this study is to focus on the relationship between shopping area image and patronage behavior by taking into account the need to develop an appropriate shopping area image scale and to ensure that shoppers are familiar with the area before they are asked about its image. To this end, data was obtained from a large scale study that was designed to obtain information on the shopping behavior of consumers in London (a mit-size city with a very well-developed public transportation system), Canada. The study required the respondent to complete an eight-page questionnaire that included questions on the image of the downtown shopping area, and to keep a two-week diary that recorded information on shopping trips and expenditures to that area. Based on the results of two pilot studies, it was decided that principal shoppers, who in most cases may be females, would be the subjects of this study. Thus, initial contact letters, personally addressed to the female heads of households whenever possible, were mailed to 2070 residences in the city. A telephone call followed a few days after the letter. Of the 1269 principal shoppers reached by telephone, 823 agreed to participate in the study and a total of 679 returns were received. Of these, 482 respondents completed both the questionnaires and diaries and they formed the usable data-base sample. The return rate was equivalent to 58.6% of those who agreed to participate or 38.05 of those reached by telephone. Considering the level of difficulty of the survey instruments, the return rates were considered very satisfying.

To overcome the familiarity issue, only those respondents (n=479) who hat ever been to the downtown area were considered for the analysis. The respondents represented a cross-section of residents in the city in that they did not differ from the original sampling frame by postal region. In addition, consistent with prior expectation, 91.3% of the principal shoppers were females and only 51.3% were employed. Non-response bias was determined in two ways. For those people who refused to participate in the study, three demographic questions on age group, educational level and years living in the city were asked over the telephone. The results of the difference of means test of the nonparticipants against the data-base sample showed that there were no significant differences with regard to the length of residence in the city and the education level. The only significant difference was that of age. Considering the nature of this study, the results were not surprising as older people tend to have more difficulties with their seeing and writing capabilities and thus tend to shun away from participating (Dillman 1978, p.53). In fact, the refusal rate was 42.1% for those people over 55 years of age.

As this study involved completion of both the questionnaire and diary, non-response bias was further assessed between those respondents who completed only the questionnaire versus the data-base sample. Comparison between these two groups were made along 7 criteria -- age, education, Siegel's (NORC) job prestige scale, income, years of living in the city, years living at the present address, and the number of shopping areas visited over the last three months. The only significant difference, using the difference of means test, between these two groups, at the conventional 5% level of significance, was that of educational level. Again, this result was not unexpected. It was possible that those participants with lower level of education had more difficulty in responding to both the questionnaire and diary.

Taking into account that there were no significant differences within each of the two sets of comparisons that were discussed above, it was concluded that nonresponse bias was not a serious problem in this study.


To begin with, the number of items in the image scale were developed from several sources. A list of items was first drawn from a literature review of store and shopping area image research. Three "experts", consisting of the author, another doctoral student, and a professor then discussed and evaluated the items. The list of items were then pilot tested on a group of 6 retailers in the downtown area and a group of 8 consumers (6 females and 2 males). A focus group interview was conducted where among other items in the questionnaire, the image scale items were discussed among the 8 consumers and the three experts. This led to editing, purification and improvement in the scale. The improved scale of items was then tested on the same 6 retailers and another 7 female consumers. Another consumer focus group interview was conducted and the scale was purified and edited further. This finally resulted in a list of 31 items for the image scale which could be considered to have content and ,ace validity, or intrinsic validity (Nunnally 1978, p.91-94 Guilford 1954, p 400). It is also important to point out that a 5-point Likert type scale was adopted after taking into account various factors that could affect a scale construction (Andrews 1982), and that 15 of the 31 items were stated negatively to prevent response style bias.

Upon collection of data, the domain sampling model (Nunnally 1978, p.193-200) was used to purify the scale further. However, as all the items were highly intercorrelated, no item was eliminated from the scale. The internal consistency of the items in the image scale was further checked by the Cronbach's coefficient alpha (Cronbach 1951) which was 0.82.


Factor analysis was next carried out on the image scale to determine the component structure. R-factoring with principal component analysis was used and the obtained factors were rotated using the varimax method. Using the criterion of meaningfulness advocated by Gorsuch, and also shared by Cattell (1979, p.485), those items that did not have a factor loading of at least 0.30 on any factor were eliminated. In addition, those items that loaded heavily on more than one factor were also eliminated (Gorsuch 1983, p.210; Churchill 1979). Applying these two criteria resulted in 4 items been removed from the scale, leaving 27 items.

With regard to the number of factors to be extracted, two criteria were considered. Using the roots criterion (eigenvalues greater than 1), a maximum of 9 factors would be extracted. However, the roots criterion is only accurate when the number of variables is small to moderate and the communalities are high. When large number of variables are involved, the roots criterion becomes inaccurate (Gorsuch 1983, Ch.8; Stewart 1981). A better procedure for extracting factors is that of Cattell's scree test (Cattell 1979, Ch.5). The scree plot, however, did not provide a very clean break in the "elbow", and suggested a possible 3 or 4 factors. However, taking into account that the final conclusions of a factor analysis are likely to be more distorted by underfactoring than over-factoring (Cattell 1979, p.55), 4 factors were extracted (see Figure 1).





The newly identified component scales are shown in Table 1. The first factor contained 7 items that were related to the total product and service offerings of the area, and was accordingly named assortment factor. The second factor had 9 items that were related to making the area more attractive to the shopper by its supportive facilities; that is, they were variables that helped facilitate the competitiveness of the area. Accordingly, it was named facilitative factor. The third set of 7 items were related more to maintaining the competitiveness of the area, and hence vas appropriately called maintenance factor. Finally, the last set of 4 items were related to the day-to-day operations of the area, and was named operational factor. She coefficient alphas of these four factors ranged from 0.58 to 0.74.


The best way to test construct validity would have been the multi-trait, multi-method matrix (Campbell and Fiske 1959). However, available resources did not permit the application of this method in the design and development of the instruments. Another acceptable way to test the adequacy of the outline of a domain relating to a construct is to determine how well the measures of observables "go together" in empirical investigations (Nunnally 1978, p.100). In this case, the construct validity of the image scale can be tested empirically by determining how well the item scores correlate with one another. Thus if all the measures purportedly measure the same construct, they should be highly correlated or in Blalock's (1968) term, an "epistimic correlation." In addition to high correlations, the series of items should also display internal consistency. Coefficient alpha is useful for this purpose because it provides both a summary measure of the homogeneity of a set of variables and an estimate of the reliability of alternative forms of the instrument.



Table 2 provides very strong support that the image scales used in this study had considerable construct validity. Overall, the high inter-item correlations were reflected by the very low value in the determinant of the correlation matrix of 0.0051. This was further supported by the fairly high item-total correlations and the respective squared multiple correlations, where the majority of the values was over 0.30. The coefficient alpha for the final total scale was 0.82 giving credence to support the fact that all the image items were related to a single underlying construct.


The summated scores of each of the four components of the image construct -- the assortment factor (7 items), the facilitative factor (9 items), the maintenance factor (7 items) and the operational factor (4 items) -- were then used to determine the relationship between image and shopping behaviour. Stepwise regressions were used and the results are shown in Table 3. Only two out of four factors - the operational and facilitative factors - were significantly loaded in the final equation for both the trip and expenditure data. For the trip data, the operational and facilitative factors accounted for an adjusted R-square of 0.162. The same two factors accounted for an adjusted R-square of 0.118 for the expenditure data.



In terms of explanation, the operational factor was clearly the more dominant factor as it entered the regression equation first in both cases and also had a much higher beta value. It was at least 2.5 and 2.0 times as great as the effect of the facilitative factor for the trip and expenditure data respectively. In both regression equations, the two factors were highly significant as indicated by the T-values.


In terms of the image construct, it is interesting to note that the assortment and facilitative factors of the image scale contain items that are very similar to those of Nevin and Houston (1980) study. However, the maintenance and operational factors contain very different items. The results confirm remarks made by Lindquist (1974-75) and other researchers about the complexity of the image construct. More importantly, the maintenance and operational factors in this study contain items that tend to reflect more the characteristics of a shopping area, and underline the importance of sampling a wider domain of items in a shopping area image study.

What has made earlier image research practical, acceptable and useful is that the findings have been applied to the positionings and development of marketing strategies by stores and shopping centers. The underlying assumption is that image can affect patronage behavior. This study provided the empirical evidence to support the impact of image on shopping trips and expenditure. lt also confirmed the predictive validity of the shopping area image scale that was developed for this study. What was interesting, however, was the type of image components that were significant or insignificant in predicting and explaining shopping behavior. The assortment factor which represented the "product and service offering" of the area and the maintenance factor were both not significant in affecting patronage behaviour. There are intuitive reasons for the lack of affects of these two factors. In general, shoppers may take for granted that for a large shopping area like downtown, it has to have a large assortment of stores and services. Therefore, because of this expectation, the assortment component of the image scale will not become a discriminating factor for visiting the area. Similarly, the downtown is generally the "heart" of the city, and shoppers may expect that it should be well kept and maintained. As such, this could possibly explain for the lack of effects of the maintenance factor.

The two image components that significantly affected shopping behaviour were operational and facilitative factors. In particular, the operational factor was at least twice as important as the facilitative factor in explaining patronage behaviour. Therefore, improving the operational factor should be an important element of a downtown strategy. This can include extending the store hours as well as providing some protection against the elements of nature. For example, building canopies along sidewalks may be a good approach to take. In the long run, it may be worthwhile to consider providing internal linkages within the whole downtown core so as to enable it to be marketed as a truly integrated shopping area.

Many of the items within the facilitative factor relate to making downtown a more attractive shopping area. Creating a more congenial environment for shopping, such as providing more common facilities like washrooms and children's play areas should help increase the level of patronage. In particular, there were three items representing parking. Improving the parking situation should help the business in downtown. This can be done in many ways. For example, merchants can help absorb part or all the parking costs of customers who made purchases at the stores. Lobbying the City officials for more parking spaces in downtown as well as extending the time on street parking meters may help. More importantly, there should be enough parking spaces near to where customers are shopping. In addition, in any future expansion or renovation in downtown, parking should be a key consideration in any building plans.


This study is not without weaknesses. By focusing only on one shopping area, it is impossible to infer what attributes determine a consumer's choice among various shopping centers. Although it has been shown that there is a relationship between patronage behavior and the image of a particular shopping area, it is difficult to know if the presence/absence of other shopping areas (and their images) would have affected the results.

Image, in terms of the operational and facilitative factors, only accounted for 16% and 12% of the variance of shopping trips and expenditure respectively. There remains a large proportion of variance not explained by the image construct. Variables like the size of the shopping area, the distance of the shopper's home to the shopping area, the presence or absence of large well-known stores, travel mode attributes, and other situational factors may be useful predictors to include in future studies. In fact, some studies have shown that the size of the shopping area and the distance to it can affect patronage behavior (for example, Gautschi 1981; Wee and Pearce 1985).

There are several ocher questions that remain to be answered with regard to research on retail images. For example, to what extent does the image of a store influence the image of a shopping center, and vice versa? So far, there have been no study on this. Does shopping center image influence some consumers more than others? Within the same consumer, does shopping center image influence him/her more in some situations than others? In other words, does image have situational effects? Finally, do consumers form similar images of different stores (branches) that belong co the same company, that is, can image be transferable? These are worthwhile avenues for future research.


Acito, F. and R.D. Anderson (1979), "The Influence of Learning on Store Image Differentiation: An Empirical Demonstration," in N. Beckwith et al (eds), Educators" Proceedings, Chicago: American Marketing Association, 500-03.

Andrews, F.M. (1982), "The Construct Validity and Error Components of Survey Measures: Estimates from a Structural Modeling Approach," Working Paper Series (May), Institute for Social Research: University of Michigan.

Bellenger, D.N., D.H. Robertson and B. Greenberg (1977), "Shopping Center Patronage Motives," Journal of Retailing, Vol. 53 (Summer), 28-38.

Berman, B. (1983), "Shopper and Retailer Perceptions of the Relative Strengths and Weaknesses of Sixteen Business Districts," Spring Conference (April 16), Presentation to the American Collegiate Retailing Association.

Blaclock, H.M. (1968), "The Measurement Problem: A Gap Between the Language of Theory and Research," in H.M. Blalock and A.B. Blalock (eds), Methodology in Social Research, New York: McGraw Hill, 5-27.

Campbell, D.T. and D.W. Fiske (1959), "Convergent and Discriminant Validation by the Multitrait-Multimethod Matrix," Psychological Bulletin, Vol. 56.

Carter, C. (1978), "Why They Shop Some Centers," Chain Store Age Executive, (May), 31-35.

Carter, C. (1981), "Why They Shop Some Centers," Chain Store Age Executive, (September), 57-59.

Cattell, R.B. (1979), The Scientific Use of Factor Analysis in Behavioural and Life Sciences, New York: Plenum Press.

Churchill, G.A., (1979), "A Paradigm for Developing Better Measures of Marketing Constructs," Journal of Marketing Research, Vol. 16 (February), 64-73.

Cronbach, L.J. (1951), "Coefficients Alpha and the Internal Structure of Tests," Psychometrika, Vol. 16 (September), 297-334.

Dillman, D.A. (1978), Mail and Telephone Surveys: The Total Design Method, New York: Wiley.

Frederick, C., D. Norton and M.L. Shannon (1975), "Tuscaloosa CBD Redevelopment Feasibility Study," unpublished report, University of Alabama; reported in J.B. Mason and M.L. Mayer (1978), Modern Retailing: Theory and Practice, Dallas, Texas: Business Publications Inc., p.103.

Gautschi, D.A. (1981), "Specification of Patronage Models for Retail Center Choice," Journal of Marketing Research, Vol. 18 (May), 162-74.

Gentry, J.W. and A.C. Burns (1977-78), "How Important are Evaluative Criteria in Shopping Center Patronage," Journal of Retailing, Vol. 53 (Winter), 73-86.

Gorsuch, R.L. (1983), Factor Analysis, 2nd Edition, Hillsdale, New Jersey: Lawrence Erlbaum Associates.

Guilford, J.P. (1954), Psychometric Methods, 2nd Edition, New York: McGraw Hill.

Hauser, J.R. and F.S. Koppelman (1979), "Alternative Perceptual Mapping Techniques: Relative Accuracy and Usefulness," Journal of Marketing Research, Vol. 16 (November), 495-506.

Hirschman, E.C. (1981), "Retail Research and Theory," in B. Enis and K. Roering (eds), Review of Marketing, Chicago: American Marketing Association. 120-33.

Houston, M.J. and J.R. Nevin (1980), "Retail Shopping Area Image: Structure and Congruency Between Downtown Area and Shopping Centers," in Rent B. Monroe (ed), Advances in Consumer Research, Vol. 8, Ann Arbor: Association for Consumer Research, 677-81.

Howell, R.D. and J.D. Rogers (1980), "Research into Shopping Mall Choice Behaviour," in Kent B. Monroe (ed), Advances in Consumer Research, Vol. 8, Ann Arbor: Association for Consumer Research. 671-76.

Lindquist, J.D. (1974-75), "Meaning of Image: A Survey of Empirical and Hypothetical Evidence," Journal of Retailing, Vol. SO (Winter), 29-38.

Moore, Charles T. and J. Barry Mason (1969), "A Research Note on Major Retail Center Patronage," Journal of Marketing, Vol.33 (July), 61-63.

Nevin, J.R. and M.J. Houston (1980), "Image as a Component of Attraction to Intraurban Shopping Areas," Journal of Retailing, Vol. 56 (Spring), 77-93.

Nunnally, J.C. (1978), Psychometric Theory, 2nd Edition, New York: McGraw Hill.

O'Neill, M.F. and D.I. Hawkins (1980), "Managerial Bias in Anticipated Images in Competing Shopping Areas," Journal of Business Research, Vol. 8, 419-28.

Stewart, D.w. (9181), "The Application and Misapplication of Factor Analysis in Marketing Research," Journal of Marketing Research, Vol. 18 (February). 51-62.

Wee, Chow Hou (1985), "Complexity of Shopping Area Image: Its Factor Analytic Structure in Relation to the Effects of Familiarity and Size of the Area," in N.K. Malhotra (ed), Developments in Marketing Science, Atlanta: Academy of Marketing Science, Vol. VII, 196-201.

Wee, Chow Hou and M.R. Pearce (1985), "Patronage Behavior Toward Shopping Areas: A Proposed Model Based on Huff's Model of Retail Gravitation," in E.C. Hirschman and M.B. Holbrook (eds), Advances in Consumer Research, Provo, UT: Association For Consumer Research, Vol. 12, 592-597.



Chow Hou Wee, National University of Singapore


NA - Advances in Consumer Research Volume 13 | 1986

Share Proceeding

Featured papers

See More


Flavor Fatigue: How Cognitive Depletion Reduces Enjoyment of Complex Flavors

Rhonda Hadi, Oxford University, UK
Dan Rubin, St. John’s University
Diogo Hildebrand, Baruch College, USA
Thomas Kramer, University of California Riverside, USA

Read More


Thank You = Trust Me: When Gratitude Expressions Help Promote New Products

Maria Ng, University of Houston, USA
Dejun Tony Kong, University of Houston, USA
Vanessa Patrick, University of Houston, USA

Read More


Tattoo: A Perspective Beyond Estethics

Luana C. Moraes, Universidade de Sao Paulo
Gabriela L. Pinheiro, Universidade de Sao Paulo
Nathalia S. Arthur, Universidade de Sao Paulo
Eliani C. Flores, Universidad Catolica del Peru
Jose Mauro C. Hernandez, Centro Universitário FEI

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.