The Consumer Expenditure Survey: Prospects For Consumer Research

Gregory. D. Upah, Virginia Polytechnic Institute and State University
Seymour Sudman, University of Illinois
ABSTRACT - This paper provides a brief overview of methodological changes in, and particular benefits of, the Consumer Expenditure Survey as they relate to prospects for consumer research. The major focus of the paper is to point out some of the major issues/research questions which can be addressed through the use of the C.E.S. data.
[ to cite ]:
Gregory. D. Upah and Seymour Sudman (1981) ,"The Consumer Expenditure Survey: Prospects For Consumer Research", in NA - Advances in Consumer Research Volume 08, eds. Kent B. Monroe, Ann Abor, MI : Association for Consumer Research, Pages: 262-266.

Advances in Consumer Research Volume 8, 1981      Pages 262-266

THE CONSUMER EXPENDITURE SURVEY: PROSPECTS FOR CONSUMER RESEARCH

Gregory. D. Upah, Virginia Polytechnic Institute and State University

Seymour Sudman, University of Illinois

ABSTRACT -

This paper provides a brief overview of methodological changes in, and particular benefits of, the Consumer Expenditure Survey as they relate to prospects for consumer research. The major focus of the paper is to point out some of the major issues/research questions which can be addressed through the use of the C.E.S. data.

As early as 1965, there were major attempts to stimulate increased interest in longitudinal and other types of overtime studies of consumer behavior (Andreasen 1965; Granbois and Engel 1965; Nicosia 1965). However, to date there have been few such efforts undertaken and reported in the literature by consumer researchers. While such research is widely done by industry sponsored research groups (e.g. The Conference Board), major marketing research organizations and major retail and manufacturing firms, seldom are these results published and/or available to consumer researchers not associated with these organizations (for exceptions, see Linden 1965 and Linden and Axel 1978).

The Consumer Expenditure Survey sponsored by the Bureau of Labor Statistics and conducted by the Bureau of the Census provides in many ways the most comprehensive over-time data on consumer behavior of any such "panel" study in existence. This data base has been frequently utilized by home economists and consumption economists. However, this data and similar data collected in other countries have received only scant use by, and attention from, consumer researchers.

As Sheth (1980) points out, much existing government and private data on consumer behavior goes unanalyzed, while more data is collected; more theories are left untested while new ones are developed. Furthermore, much of this government data is collected from large, national probability samples--just the sort of data base which is generally not economically feasible for most academic research. Sheth suggests that the analysis of existing data become a priority for future consumer research efforts. The existence of the Consumer Expenditure Survey data provides one major opportunity for those interested in any number of aspects of consumer behavior to do just this.

PURPOSE

The purpose of this paper is to: (1) provide background information on the methodological changes in the Consumer Expenditure Survey, (2) compare and contrast the current C.E.S. with other well-known panel studies, (3) outline the unique advantages and potential drawbacks of the C.E.S. for consumer research and (4) suggest several sets of illustrative research questions for which C.E.S. data should be particularly appropriate for answering.

BACKGROUND

As has been discussed in other papers in this session, the current methodology used for the current C.E.S. is markedly different from that which was used in the previous surveys done in 1950 and 1960. In those years, respondents were asked, in basically one or two long interviews, to recall all of their household expenditures throughout the previous year. There is substantial evidence that this task was too difficult for many respondents, particularly for non-durable goods. The recall method has, even for shorter recall periods, been shown to produce seriously inaccurate reporting (Sudman and Ferber 1974).

Beginning with 1972-1974 survey, and continuing through the current survey, the diary method has been used to collect expenditures on inexpensive non-durables. The recall method is used to collect expenditure data for more expensive durable goods and regular expenses such as rents, utility bills and insurance premiums. The recall period for durables is 3 months. Diary data is collected only over a 2 week period.

BLS is now engaged in an ongoing or continuous Consumer Expenditure Survey Program. Whereas the 1950,1960 and 1972-74 surveys were one-time surveys, the current survey collects data on at least a monthly basis. The survey's dynamic character, along with its improved data quality, allows for analyses not possible with previous surveys. For instance, analyses of the short-term impacts of major social events could not be made with the 1950 and 1960 surveys since the data collected was for a single one-year period. While such analyses--over a 15 month period were possible with the 1972-74 data, measures of expenditure changes in the years after 1961 and before 1972 and in the years after 1972 and prior to the current survey were not.

The data quality especially for non-durable goods purchases has been markedly improved (since the 1960 survey) through the use of the diary method. The data on these expenditures may be more comprehensive than any such data previously collected. As Pearl (1981) points out, the Consumer Expenditure Survey obtains the most extensive travel expenditure data ever collected.

SOME MAJOR COMMONALTIES AND DISTINCTIONS BETWEEN THE C.E.S. AND COMMERCIAL PANEL STUDIES

There has been somewhat substantial use of and reference to consumer panel data in the marketing literature (e.g., Ahl 1970; Cherington 1943; Frank and Strain 1972; Massy and Frank 1965). For a review of the use of consumer panels in marketing see Sudman and Ferber (1979). Some of the same objectives illustrated in these studies can be satisfied through use of the C.E.S. data. A comparison of the C.E.S. to a major consumer panel such as Market Research Corporation of America (MRCA) national consumer panel, does however, reveal major similarities and differences. Both "panels" provide over-time data on the quantity and type of item purchased. Both panels provide data on prices paid.

The C.E.S. does not, however, provide information on the particular brand purchased or whether or not the item was purchased with a coupon or some other promotional "deal." The latter types of data are usually obtained in commercial panel studies. The C.E.S. does not obtain data on outlets shopped. However, a separate study sponsored by BLS, the Point-of-Purchase Survey, does provide a means for some types of outlet studies.

Some commercial panels also have measures of a wider variety of household characteristics to relate to expenditures--e.g., life styles, than are obtained in the C.E.S., as well as attitudinal measures regarding products for which purchase histories are obtained. The Consumer Expenditure Survey obtains no data on attitudes, life styles or behavioral intention. Thus, while desirable, over-time analyses of the relationships between these other household or individual characteristics and behavior are not possible with existing C.E.S. data. However, as Jacoby (1978) points out, the analysis of consumption itself and its consequences for subsequent acquisition decisions deserves more attention in the consumer behavior literature.

The C.E.S. and MRCA panels differ in terms of the length of time participants remain in the panel. The maximum length of time for participation in the C.E.S. is 15 months. MRCA's panel members may, for instance, remain in the panel indefinitely. This continual replacement effectively removes the C.E.S. from categorization as a longitudinal study. Longitudinal analysis requires over-time observations to be made on the same sampling units. The C.E.S. is longitudinal for a maximum of 15 months--the five quarters for which data is collected. However, no longitudinal analysis is possible for non-durables since this data is collected for only one two-week period.

Perhaps the major difference between the C.E.S. and other major panels is the breadth of the consumption expenditures measured for each sampling unit. MRCA has a panel for grocery items and another for clothing. General Electric has a panel which is concerned with major appliances. The C.E.S. measures virtually every conceivable type of expenditure using the same panel. The existence of this almost totally comprehensive expenditure data provide the unique opportunity for assessing expenditures on certain goods and services relative to expenditures on other products. This extensive detail is necessary to allow for the continual updating of weights and market basket items used in calculating the Consumer Price Index (CPI). Within some limits to amount of detail in, and specificity of, product breakdowns researchers interested in consumer expenditures on virtually any type(s) of product(s) can obtain high quality data through use of the C.E.S.

Some commercial research firms have been amenable to releasing panel data to academic researchers a specified number of years after the data has been collected. Current panel data is however, for most academic researchers, too costly. In contrast, the data tapes sold by BLS are relatively inexpensive.

Sales and distribution audits also provide data on consumer expenditures for a variety of products. Information on expenditures across types of retail outlets also is available. However, these data offer no opportunity for analyses of differences in expenditures by various subgroups of the population. An additional drawback is the relatively high cost of current data.

OTHER BENEFITS OF THE CONSUMER EXPENDITURE SURVEY DATA

As mentioned above, the C.E.S. provides interested researchers with the opportunity to analyze over-time data collected from a large national probability sample. There are some methodological concerns regarding the validity and reliability of some of the data (see other papers in this session--e.g., Pearl 1981). Nevertheless, the data quality, overall, is likely to be superior (i.e., in terms of producing lower total survey error on any given item of data) to most data which has been utilized in academic research (especially those data derived from student or local household convenience samples).

The C.E.S. data is particularly well-suited to macro or group level, as opposed to micro, analyses of consumer behavior. The former approach has been classified as a major need for consumer research in the 1980's (Sheth 1980). The sheer comprehensiveness of expenditure data present in the C.E.S. allows researchers to overcome the weaknesses in analyzing consumption of a single product and making generalizations about consumer behavior based on that single category of product. The C.E.S. data allows for an analysis of the total consumption structure of a household.

The unit of analysis--the sampling unit for the C.E.S. data, called the consumer unit, is almost always the household. The view that the household, as opposed to the individual, is the appropriate unit of analysis for consumer research has gained growing support (Ferber 1980). Only in certain circumstances (i.e., based on the financial independence of one cohabitant from another cohabitant) will a household--i.e., place of residence, contain more than a single consumer unit. BLS estimates that 95% of all households will contain a single consumer unit.

Although the C.E.S. and commercial panel data are based on self-report data as opposed to direct observation or record checks, the data deals with actual purchases--i.e., expenditures. There are no attitudinal measures taken.

Kassarjian, for one, has argued for more research into consumer behavior which relies on unobtrusive and nonattitudinal measures--(e.g., analysis of current periodicals, advertisements, etc.) insofar as they reflect the changing values and attitudes of a society. Actual expenditure data deals with the end-result or activity that an analysis of these factors intends to predict. Researchers who wish to make inferences as to what value/attitude changes are taking place can look to the current production of visible tangible symbols of a society (its art, literature, etc.). However, it also is important to analyze the actual consumption patterns of various segments of the society and society in general. There are, of course, relative disadvantages to any single approach--e.g., determining whether or not changes in art, or literature are lagging behind or preceding the value or attitude changes that help to produce changes in consumer expenditures. Furthermore, consumer expenditures are constrained by what sellers choose to offer in the marketplace. Thus, consumer behavior may be reflective not of what consumers are thinking, feeling or desiring, but rather what compromises from what is wanted or needed they are willing to make given the limits of the marketplace.

The complex and interesting questions surrounding the macro aspects of consumer behavior render any one means of assessing long term shifts in value and expenditure somewhat inadequate. Multiple means/indicators are required. Nevertheless, the comprehensive expenditure data available through the C.E.S. is one such means of assessment that has not been fully utilized by consumer researchers.

POSSIBLE RESEARCH PROGRAMS UTILIZING THE C.E.S. DATA

The following are illustrative examples of the research questions and issues for which the use of the C.E.S. data would be most appropriate and useful.

1. Analysis of Long Term Predictions for Changes in Consumer Behavior:

How do changes in consumer expenditures compare to the long-term changes in consumer behavior predicted by various pundits and forecasters?

For instance, the Yankelovich research organization regularly tracks changes in a number of major cultural values. Based on this organization's surveys, predictions for major changes in consumer values in the U.S. are made. These changes in values will ultimately influence attitudes and purchase behavior toward a wide variety of goods and services (Kotler 1980: Loudon and Della Bitta 1979). The Consumer Expenditure Survey data can be used as one means of testing specific propositions about changing values and consumer behavior through an analysis of changes in consumer expenditures.

2. How are consumer expenditure patterns influenced by the major changes in characteristics of households?

(a) How do consumption expenditures of married households in which both spouses work differ from those married households with wives who are not employed outside the home? This is a popular issue and source of conjecture among home economists, consumer researchers and marketers. Still, there have been few empirical studies examining the full range of implications of the working wife on household buying behavior.

One study dealing with this issue utilized data obtained from a 1967-70 Survey Research Center study of consumer finances. Holding income, education, family life cycle and other socioeconomic and demographic measures constant, the labor-force participation of wives was related to overall household consumption expenditures, and expenditures on durable goods (both as a percent of total income) (Strober 1977). Strober found that holding other factors constant, households with working wives had a significantly higher consumption expenditure to income ratio than did households without working wives. The labor-force participation of the wife, however, had no significant effect on total expenditures or durable goods (as a percent of total household income). This study tests several hypotheses developed in economics regarding the expenditures of households with working wives (see Strober 1977 for a brief overview of these hypotheses). There seems to be a great opportunity to merge the consumer behavior theory developed in sociology, marketing and psychology with these economic theories to further explore these issues. At the very least additional tests of the hypotheses developed in economics could be made. The Consumer Expenditure Survey data again provides an excellent data base for such research.

(b) How do these consumption expenditure patterns change as a function of the age of the wife/husband, number and ages of children or overall family life cycle of the household?

There have been two recent studies of the impact of family life cycle on various consumer expenditures. The first of these studies was, in fact, based on Consumer Expenditure Survey data collected by the Government of Norway (Arndt 1979). The other study was based on a local commercial survey in the U.S. (Landon and Locander 1979). Both of these studies have, however, been criticized for failing to allow for, or to hold constant, other factors (aside from family life cycle) likely to influence the expenditures under consideration--e.g., income, education and occupation (Ferber 1979). As such they provided no opportunity to assess the efficacy of family life cycle relative other household characteristics in explaining consumer expenditure.

The Consumer Expenditure Survey provides ample data for constructing various formulations of family life cycle. Those interested in conducting such research might veil consider the criticisms, and suggestions for modification, of the family life cycle concept itself (Murphy and Staples 1979). In addition, the Consumer Expenditure Survey data contains measures of the household characteristics needed for the multivariate analyses required to establish the usefulness of the family life cycle concept--in whatever form, for explaining consumer behavior. The continuous nature of the survey allows for an analysis of the short-term impact (i.e., impacts on durable goods expenditures) of changes in household composition or life cycle--e.g., a new baby.

(c) Assessing the impact of other changes in household composition.

The dynamic character of the current C.E.S. allows for an analyses of the short and long-term effects of the aforementioned changes in household composition. The effects of other changes such as a loss of job or death in the family also can be assessed.

3. How do consumption expenditure patterns of Blacks, Whites and Hispanics differ? Do these expenditures differ when age, education, income and other relevant demographic characteristics are held constant or controlled for across the two groups? If so, which types of expenditures differ?

Recent research indicates that in similar occupational, income, educational age and life style categories there are few Black-White differences in consumption (Edmonds 1979). The Consumer Expenditure Survey offers an extensive data base for making these cross-sectional comparisons over time.

4. What changes in consumption (e.g., percent of income spent on the good or service) occur as income increases?

These issues have concerned economists (e.g., with respect to Engel's Law) and other researchers for a number of years (see Prais and Houthakker 1971). It should be interesting to note where these shifts have been occurring in recent years.

BLS has, in past years, calculated income sensitivity ratios. These ratios represent the percent increase in expenditure on some item (i.e., usually a broad category of items) resulting from a one percent increase in real disposable income. One drawback in the use of these ratios has been their timeliness (Reynolds and Wells 1978). However, with the switch by BLS to an ongoing C.E.S., these ratios can be kept much more current. They provide a means for marketers to calculate likely changes in consumption of their products with changes in disposable income levels. They also provide a means of assessing the effects of sheer increases in disposable income on relative increases in the consumption of various products (i.e., tests of income--related consumption hypotheses.)

5.  Profiling the Heavy User:

Comparisons of heavy users of a particular category of good or service with light users are easily facilitated with the Consumer Expenditure Survey. These comparisons can be made in terms of the various household characteristics which are measured or can be derived (e.g., family life cycle) from the C.E.S. survey. Comparisons in terms of use of related goods or services across user categories also could be made. One possible research focus would be to compare the results to results of past "heavy half" studies.

6.  The Growth of Services in the Economy

Services have, over the past ten years, grown as a percent of GNP. Potential research questions regarding the growth of services include: Which particular service expenditures have exhibited the greatest increase? How do these rates of increase compare across categories of households? Are these expenditure changes reflective of recent predictions for consumer shifts to the service sector of the economy? These changes in the over-time expenditures on services are best measured by the continuous survey now in progress rather than by separate surveys conducted at yearly or 5 or 10 year intervals.

7.  Applications of Life Style or Consumption Style Approaches to the Study of Consumer Behavior

While the Consumer Expenditure Survey does not obtain life style data based on activity interest and opinion questions, it does allow for a means of inferring life styles from actual consumption behavior. The use of actual consumption data is one of several means of assessing life styles (Wind and Green 1974). The use of particular products, which might be considered to be part of the activity component of life styles, has been correlated with the purchase of other products of interest (see, for example, Hustad and Pessemier 1974--for a review of such studies, and Bass, Pessemier and Tigert 1969 and Reynolds and Wells 1978).

Sheth (1980) discusses the need for more research dealing essentially with a typology of consumption life styles. This research would involve an analysis of "the time dependent covariances of preselected and representative goods and services." The Consumer Expenditure Survey data is ideally suited to just such a research approach. Life styles can thus be inferred from actual consumption patterns and not merely activity interest and opinion data. Analysis of trends in life styles based on changes in consumption of various types of products would presumably be of major interest to marketers, consumer researchers interested in the overall consumption trends in the market place, and sociologists and others who wish to discover evidence for changes in values and life styles, over time.

There has been some recent research in these areas. One study analyzed the correlates and consumer behavior implications of changes in consumption style (Uusitalo 1980), The author determined a variety of consumption styles by forming eight combinations of three consumption indices. The consumption indices were based on factor scores produced in a factor analysis of consumer expenditures on 220 items. The expenditure data was derived from Consumer Expenditure Survey data collected by the Statistical Central Office of Finland. The consumption styles derived from these analyses were then related to total consumption expenditures and individual expenditures on specific categories of products.

One possible caution with respect to such analyses is in relating predictor variables derived from particular expenditures to essentially the same expenditures. A better method would be to derive the consumption style measures from one set of expenditures, and then using those measures to predict another set of expenditures.

8.  Cross-Cultural Analyses of Consumer Expenditures

Already in this review of possible research programs involving the use of the C.E.S. data, three different national consumer expenditure surveys have been discussed. In fact, there are nine European countries with which cross-cultural comparisons are possible. Such studies are now more easily facilitated due to the use of much more comparable expenditure survey methods (see, Statistical Office of the European Communities 1980). One additional interesting aspect of such research is the possibility for comparisons of expenditures with Iron Curtain Countries.

9.  Examining Trends in Consumption Using New Categorizations of Goods and Services

What do sports equipment, vacations, and tickets to a play have in common in terms of satisfying consumer needs? How are they representative of different types of consumer needs? What are the underlying consumption trends in the vast array of changes in consumer expenditures? Clearly, better characteristics of products (goods and services) theories--i.e., categorizations of products according to what value expression or sought need they represent are needed to answer these types of questions. The convenience goods-shopping goods-specialty goods breakdown clearly is not sufficient for making projections in needs-based categories of expenditures, or in making inferences from expenditure data to value shifts. Work in analysis of consumption as it pertains to the acquisition of particular benefits has been done by Lancaster.(1975) and Ratchford (1975).

Some trend studies of consumer expenditures published by BLS break all consumer expenditures down to expenditures on food, housing, clothing, medical care, transportation, recreation (which includes among other things expenditures on education, reading, alcoholic beverages and tobacco). There are certain to be subtle differences within each of these expenditure categories which when guided by a theory-based approach could yield new insights into consumer behavior. Furthermore, as Nicosia and Mayer (1976) point out, differences in the nature of expenditures made within a particular expenditure category may have important public policy implications. For example, the satisfaction of entertainment needs via attendance at spectator sports versus purchase of records or books will influence energy utilization, use of roads and mess transit, consumption of food and drinks away from home etc.

10. Impact of Major Social Events on Consumer Behavior

As mentioned above, the continuous collection of expenditure data allows for an assessment of the short-term impacts of major social events on consumer behavior. For instance, the impact of such events as oil embargoes and resulting gasoline shortages, the eruption of Mt. St. Helens, major new product introductions such as K-Body cars or video disc players, shortages of certain products, new laws, major economic changes (e.g., rises or falls in inflation) or political changes (e.g., the election of a new president) can be measured by the continuing survey.

CONCLUSION

The Consumer Expenditure Survey has provided, and continues to provide, interested consumer researchers a rich data base which can be used to satisfy a variety of research interests. There is certainly room for creative, theory-based approaches to the analysis of the C.E.S. data. Fundamental tabulations and cross-tabulations of expenditure data--of the type published by BLS and the Conference Board are some of the most important analyses performed. However, there is much room to go beyond these analyses to answer some of the more complex questions consumer researchers and public policy makers have, especially in recent years, been posing. The C.E.S. data holds out the promise of not only enabling consumer researchers to see what is happening and what is likely to happen in terms of actual consumer behavior in the market place, but also providing the insights and tests of hypotheses which will assist these researchers in determining the reasons for these behaviors.

REFERENCES

Ahl, David H. (1970), "New Product Forecasting Using Consumer Panels," Journal of Marketing Research, 7 (May), 160-67.

Andreasen, Alan (1965), "Potential Marketing Applications of Longitudinal Methods," in Marketing and Economic Development, Peter D. Bennett (ed.), Chicago: American Marketing Association, 261-278.

Arndt, Johann (1979), "Family Life Cycle as a Determinant of Size and Composition of Household Expenditures," in Advances in Consumer Research, 6, William L. Wilkie (ed.), 128-132.

Bass, Frank M., Edgar A. Pessemier and Douglas J. Tigert, (1969), "Complementary and Substitute Patterns of Purchasing and Use," Journal of Advertising Research, 9, p. 19.

Cherington, Paul T. (1943), "New Economic Patterns Found by Consumer Panels: Introduction," 8 Journal of Marketing (July) 41.

Edmonds, Linda L. (1979), Clothing Buying Practices and Life Style Differentials Between Employed Black and White Women, unpublished Ph.D. Dissertation, Virginia Polytechnic Institute and State University (April).

Ferber, Robert (1979), "Comments on Life Cycle Analysis," in Advances in Consumer Research, 6, William L. Wilkie (ed.), 146-148.

Ferber, Robert (1980), "What Do We Know About Consumer Behavior,'' in Marketing Management and Strategy, Philip Kotler and Keith Cox (eds.), Englewood Cliffs: Prentice Hall, 97-108.

Frank, Ronald E. and Charles Z. Strain (1972), "A Segmentation Research Design Using Consumer Panel Data," Journal of Marketing Research, 9 (November), 385-90.

Granbois, Donald H. and James F. Engel (1965), "The Longitudinal Approach to Studying Marketing Behavior," in Marketing and Economic Development, Peter D. Bennett (ed.), Chicago: American Marketing Association, 208-221.

Hustad, Thomas P. and Edgar A. Pessemier (1974), "The Development and Application of Psychographic Life Style and Associated Activity and Attitude Measures," in Life Styles and Psychographic, William D. Wells (ed.), Chicago: American .Marketing Association, 97-126.

Jacoby, Jacob (1978), "Consumer Research: A State of the Art Review," Journal of Marketing, (April), 87-96.

Kotler, Philip (1980), Marketing Management: Analysis, Planning and Control, 4th ed., Englewood Cliffs: Prentice Hall, 122-128.

Lancaster, Kelvin (1975), "Hierarchies in Goods Characteristics Analysis," in Advances in Consumer Research, Vol. 3, Beverlee B. Anderson (ed.), Ann Arbor: Association for Consumer Research, 348-352.

Landon, E. Laird Jr. and William B. Locander (1979), "Family Life Cycle and Leisure Behavior Research," in Advances in Consumer Research, 6, William L. Wilkie, (ed.), 133-138.

Linden, Fabien (1965), Expenditure Patterns of the American Family, New York: The Conference Board.

Linden, Fabien and Helen Axel (1978), Consumer Expenditure Patterns, Vol. I., New York: The Conference Board.

Loudon David J. and Albert J. Della Bitta (1979), Consumer Behavior: Concepts and Applications, New York: McGraw-Hill, 120-132.

Massy, William F. and Ronald E. Frank (1965), "Short Term Price and Dealing Effects in Selected Market Segments," Journal of Marketing Research, 2 (May), 171-85.

Murphy, Patrick E. and William I. Staples (1979), "A Modernized Family Life Cycle," Journal of Consumer Research, 6, (June), 12-22.

Mitosis, Francesco M. (1965), "Panel Designs and Analyses in Marketing," in Marketing and Economic Development, Peter D. Bennett (ed.), Chicago: American Marketing Association, 222-243.

Mitosis, Francesco M. and Robert N. Mayer (1976), "Toward a Sociology of Consumption," Journal of Consumer Research, 3, (September), 65-75.

Pearl, Robert (1981), "An Evaluation of the 1972-73 Consumer Expenditure Survey Data," in Advances in Consumer Research, Vol. 8, Kent B. Monroe (ed.).

Prais, S. J. and H. S. Houthakker (1971), The Analysis of Family Budgets, Cambridge: The University Press.

Ratchford, Brian T. (1975), "The New Economic Theory of Consumer Behavior: An Interpretive Essay," Journal of Consumer Research, 2, 65-75.

Reynolds, Fred D. and William D. Wells (1977), Consumer Behavior, New York: McGraw-Hill, 385-412.

Sheth, Jagdish N. (1979), "The Surpluses and Shortages in Consumer Behavior Theory and Research,: Journal of the Academy of Marketing Science, 7, (Fall), 414-427.

Strober, Myra (1977), "Wives' Labor Force Behavior and Family Consumption Patterns," The American Economic Review, 7, (February), 410-417.

Sudman, Seymour and Robert Ferber (1979), Consumer Panels, Chicago: American Marketing Association.

Sudman, Seymour and Robert Ferber (1974), "A Comparison of Alternative Procedures for Collecting Consumer Expenditure Survey Data for Frequently Purchased Products," Journal of Marketing Research, 11 (May), 128-135.

Uusitalo, Liisa (1980), "Identification of Consumption Style Segments on the Basis of Household Budget Allocations,'' in Advances in Consumer Research, 7, Jerry C. Olson (ed.), 451-459.

Wind, Yoram and Paul E. Green (1974), "Some Conceptual Measurement and Analytical Problems in Life Style Research," in Life Styles and Psychographics, William D. Wells (ed.), Chicago: American Marketing Association, 97-126.

Womer, Stanley (1944), "Some Applications of the Continuous Consumer Panel," Journal of Marketing, 9, (October), 132-136.

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