Advances in Consumer Research Volume 13, 1986 Pages 403-408
TOWARD THE CONSTRUCT OF CONVENIENCE IN CONSUMER RESEARCH
Laura Yale, University of California, Irvine
Alladi Venkatesh, University of California, Irvine
The authors evaluate the notion of convenience as it appears in the existing marketing/consumer behavior literature and explore the multidimensionality of the concept. With the emergence of consumer services in the past decade and the growing movement of the American society toward a service economy, the need arises to investigate consumer convenience in a systematic way. This paper addresses the issue by providing a background analysis of how the term "convenience" has appeared in marketing/consumer behavior literature and what further refinements are needed to define it in the total context of consumer behavior.
The idea of convenience was assumably first introduced into marketing by Copeland in 1923 in his HBR article suggesting the now-classic product typology: convenience, shopping, and specialty goods. His original taxonomy has been criticized and modified over the years but the idea of a convenience good as one which requires little if any cognitive decision making has survived. Recently, convenience has been empirically operationalized by various researchers (Douglas 1976; Strober and Weinberg 1977; Strober and Weinberg 1980; Schaninger and Allen 1981; Reilly 1982), but the meaning attached to the term is not that of Copeland or his successors but of time-buying or time savings instead.
The underlying hypothesis of most of these studies has been that the more constraints on a consumer's time, the more likely he/she will be to use convenience products or services, i.e., those products and services that will save or buy time. In a recent paper, however, Venkatesh and Vitalari (1985) demonstrated that not all household technologies lead to net time savings. While the theory (of time savings) is intuitively appealing it has not proven to be strongly supported. Most studies yield insignificant results or only weak Support for the hypothesis and the researchers express surprise, and some dismay, at these counterintuitive results.
In this paper, the authors propose some rationale for these enigmatic findings. It is proposed that the research suffers from three problems: 1) insignificant differences in the independent sets, 2) imprecise, "fuzzy" delineation of the dependent sets, and 3) most importantly, the lack of theoretical construction and consideration of convenience as a consumer construct.
The paper begins with a review of the original use of convenience in marketing and the theoretical underpinnings of its present use as equivalent to time savings. The recent empirical research is next summarized and critiqued. The authors then offer some preliminary theoretical development of convenience as a consumer behavior construct. Recent trends are suggested which will lead to an increase in convenience as a primary salient product attribute, and the paper concludes with proposed areas of research the authors believe are needed to develop convenience into a viable consumer construct.
Convenience Goods Perspective
Copeland (1923) defined convenience goods as "those customarily purchased at easily accessible stores..." The consumer is familiar with these articles; and as soon as he recognizes the want, the demand usually becomes clearly defined in his mind." In 1948, the AMA Definitions Committee defined convenience goods as those consumer goods which the consumer usually purchases frequently, immediately, and with a minimum of effort (Holton, 1958). Shopping and specialty goods entailed a more concerted and cognitive effort on the part of the consumer.
Holton (1958) recognized that goods vary from category to category depending on the perception and shopping propensities of the individual consumer. The consumer will consider a good a convenience (or shopping/specialty) good based upon his/her perception of the "probable gain (large vs. small) from making price and quality comparisons among alternative sellers." Interestingly, Holton suggested that nonworking wives would perceive more goods as shopping goods, while working wives would perceive more as convenience goods due to their greater opportunity cost of time. This deduction marks the link between Copeland's original usage of convenience and the recent time savings connotation.
Bucklin (1963) updated the definition of convenience goods by adding the modern concept of perceptual mapping to indicate the consumer's quick routine decision behavior: "those for which the consumer, before his need arises, possesses a preference map." Apparently, this simple definition has been accepted, by default, by marketing scholars and therefore they do not feel the need to define the term further. For example, Jolson and Proia (1976) assert "there is little disagreement as to the meaning of convenience goods." However, more recent research on the notion of convenience uses the term with a different connotation.
Convenience as a Time Dependent Construct
Through its operationalization in recent research, convenience is overwhelmingly implied to be equivalent to time saving or time buying. The theoretical framework for this new and restricted definition lies primarily in the economic concept of the household as production unit.
Becker (1965), and Michael and Becker (1973) expanded the classic economic consumer choice behavior model to include time. The classic model describes purchase behavior as a utility function in which the consuming unit attempts to maximize utility subject to constraints of income and relative prices of goods. In order to account for variations of demands (other than by the traditional economic catch-all of tastes), Michael and Becker suggest that the household produces basic commodities through the productive activity of combining purchased market goods and services with the household's own time. The household's "full income" then is composed of monetary income and total available time. Income and time constrain the production capabilities of the household.
Michael and Becker allude to the time saving connotation of recent researchers;
"it has long been recognized not only that consumers sell time in labor markets, but also that they buy tine in the form of certain consumer goods and services: (these goods and services) are all in some measure time-savers. The demand for such items would be quite different if time were not a scarce resource" (1973).
Etgar, utilizing the notion of the household as a production unit, offers the marketing theory behind the recent empirical work conducted on convenience consumption (Etgar 1978). He adds the consumption time component to previous Consumer Behavior models, transforming previous choice models into an overarching consumption model.
From the paradigm of the household as production unit, researchers, in general, hypothesize that consumers with greater time constraints will attempt to buy or save time through purchase of convenience goods (e.g., frozen dinners) and time saving durables. This intuitively appealing hypothesis, derived from the just mentioned theoretical background, has been empirically studied in the past ten years.
Findings from Previous Research
Studies, conducted within the last decade, have tested the time saving notion of convenience. However, most of these studies have focused on convenience consumption levels of working wife (WW) versus nonworking wife (NWW) households. Only one maj or study included single adult households in the study sample, and unfortunately for comparison's sake, it was merely an exploratory study (Hendrix 1984).
The empirical studies have found few significant differences in ownership of time saving durables or convenience good purchases between WW and NWW households. Strober and Weinberg (1977) considered the difference in purchase of time saving durables between WW and NWW households. Data from the 1968 Michigan Survey Research panel on Survey of Consumer Finances revealed no significant difference once income was taken into account. Mallan (1968), working with 1963-64 data, did find significant differences in WW families' likelihood to purchase durable goods. In 1980, Strober and Weinberg replicated and expanded their previous study using 1977 data and again found no significant differences between strategies used to cope with time pressures between NWW and WW households. "Convenience foods" were included as a variable in this latter study. Douglas (1976), too, found no significant difference in purchase of "convenience foods." Waldman and Jacobs (1978) also found no significant difference concerning expenditures for food away from home.
Schaninger and Allen (1981), noting the insignificant results of previous research, looked to wife's occupational status as an intervening variable to explain ownership of durables and purchase of many different products, including convenience foods. Results did indicate that, though WW and NWW households did not show a difference, wife's occupational status did explain some of the variation in consumption of convenience foods, but no significant difference in use of convenience stores and only weak support for ownership of appliances and durables.
Reilly (1982) conducted a study hypothesizing that role overload may be an intervening variable in the convenience consumption process. He found only weak support that those WW households reporting role overload consume more convenience foods and time saving durables. Interestingly, Reilly was the first empirical researcher to strongly suggest that the problem of the previous research may have been due to insufficient theoretical development of convenience as a construct:
"It may be that the measure of convenience consumption used in this and similar research is not particularly sensitive. Individuals may or may not use the measured convenience foods for a number of reasons other than a desire to save time. Similarly, the ownership of time-saving durables may be motivated by considerations other than work-load reduction. Some measurement research is needed to identify behavioral measures of convenience consumption which are not confounded with other factors.
"A second possibility is that the factors which influence the use of convenience foods and the ownership of time saving durables are not well explicated. It may be that one or more important factors were omitted here. Further research is needed to identify these factors, so that our knowledge of convenience-consumption determinants might be more complete."
The foregoing empirical studies suffer from three problems. First, the independent variables in the studies were restricted to husband-wife households. Comparisons between the stratified groups (NNW, WW, and other slightly different variations) were then made. We suggest that the stratifications (clusterings) of the groups were not dissimilar enough to allow for detection of significant differences in level of time saving durable ownership and convenience good consumption. In other words, the between treatment variation was not large enough to permit the emergence of significant differences on the dependent variables. It is possible, and we believe probable, that similar empirical studies utilizing strata representing all possible household types (i.e., singles, husband-wife, single parent, non-traditional households) would reveal significant differences in the dependent variables.
Compounding the problem of independent variable uniformity is the impreciseness in the operationalization of the dependent variables. The variables "time saving durables" and "convenience goods" were fuzzily delineated. Time saving durables included such commonly owned appliances as refrigerator, stove, freezer, waffle iron, drip coffee maker, and oven timer. Convenience goods included a wide variety of everyday products such as cream cheese spread, seasoning mixes, bottled juices, hot dogs, frozen vegetables, and teabags. Given the wide distribution and common usage of the just mentioned products, we believe these operationalizations measure the dependent variables in a very ambiguous, nonsensitive manner.
Lastly, but most importantly, the studies had a limitation because little specific theoretical development was undertaken to define the convenience construct. Few attempts have been made to reduce the ambiguities involved in the definition and operationalization of convenience. A construct can only be given clearer meaning when placed in a theoretical context which relates it to other variables. The simple unidimensionality of convenience as a time saving has proven too ambiguous for successful empirical study.
We believe it is time to pause in our empirical research and to reconsider the composition of convenience with some preliminary theoretical development. Hopefully, such theory development will lead to more fruitful operationalization of the construct.
CONVENIENCE ISSUES IN MARKETING
There are two levels at which convenience becomes an important issue in marketing: 1) the determination of convenience seeking consumer segments, 2) the determination and inclusion of convenience attributes in products and services.
Figure 1 illustrates a modeling of these two related issues.
RELATION OF CONVENIENCE ISSUES
The influencing variables are characteristics of the individual consumer. These variables will increase or decrease the consumer's preference for convenience as a consumption strategy. A high preference for convenience will lead to a heightened demand for products the consumer perceives as fulfilling his/her convenience consumption strategy. Hence, characteristics of the product and its marketing become important in motivating convenience-preferring consumer segments to purchase a product/service.
Convenience thus represents an important construct on two levels: a consumption behavior level and a product attribute level. Marketers must understand the complexity of convenience in order to determine convenience seeking consumer segments and to assign products and marketing for these segments which highlight consumer perceived convenience attributes.
PROPOSED VARIABLES INFLUENCING CONVENIENCE PREFERENCE
Etgar (1978), Feldman and Hornik (1981), Hendrix (1984), and Nicosia and Mayer (1976), each called for researchers _nd marketers alike to consider the multidimensionality of the entire household production/consumption/time allocation function. In that context, convenience, as an integral factor in consumption, should be considered in psycho-behavioral as well as simple economic and temporal terms. What follows is only an admittedly first tentative step toward theoretical development of a multidimensional construct of convenience.
By and large, convenience has been implied to involve two variables, economic and temporal. The economic dimension simply articulates the monetary value of time; that the individual can sell his time for money, or refrain from selling his time, thus incurring an opportunity cost (Mincer 1963; Becker 1965).
The temporal variable, which has been mentioned throughout the first section of this paper, also refers to the consumer's ability to buy or save time, which is strictly constrained to the total available number of hours represented by the number of members in the household times 24 hours per day. Thus both the economic and temporal variables reduce to the same thing. We believe a single temporal/economic dimension inadequately delineates convenience. One could deduce that all consumption in any trade economy can be considered convenience consumption. For example, a woman buys a suit for work because she does not have time to sew; another woman buys wool cloth to sew a suit because she does not have time to weave. Still another woman buys sheared virgin wool because she does not have time to raise sheep! Additional variables need to be considered in order to gain clearer meaning of convenience in the consumption process.
For preliminary discussion and exploration, we suggest the consideration of the following variables: spatial, psychological, sociological, philosophical, and situational.
The spatial variable is inferred from the temporal/ economic and variable. Proximity in location or in time offers utility to the consumer in a simple efficiency sense. This spatial sense is colloquially "being in the right place at the right time." It is this sense that Copeland implied in his original work. Once the consumer recognizes a need, the first opportunity to fulfill that need will be taken. The consumer will prefer to satisfy several needs in one location if given the option, as long at the economic cost incurred is not seen as significant.
There is also a psychological variable. The values, personality, opinions, attitudes, abilities, and preferences of an individual will influence his/her choice of consumption behavior. Different individuals will have different preferences for each style of consumption, each consumption strategy (Etgar 1978). A traditional mother-wife may place great value on her role as nurturer and therefore she will prefer consumption strategies which allow her to actively participate in production activities which she perceives as nurturing to her family. This same woman, when placed in a more time constrained situation (e.g., she takes a job outside the home), will most likely continue with her values and traditional consumption strategy in order to maintain her feelings of fulfillment as nurturer. Many examples such as this can be suggested to demonstrate the importance of the psychological dimension of convenience. All that will be emphasized here is that, just as psychological variables influence all individual behavior, similar variables will influence the consumer's want for, and perception of, convenience.
The construct of convenience also includes a sociological dimension, again as all human behavior does. The various roles that an individual plays will influence the individual's perception of, and use for, convenience. Reference groups such as family, church, state, and peer groups all will contribute to the importance and meaning of convenience to the individual. In addition, social class, race and cultural norms and values may suggest or preclude certain behaviors which have an impact on convenience as manifested in society. The classic story of the rationale for the exclusion of powdered eggs from cake mixes serves as an example of cultural norms influencing socially acceptable convenience-seeking behavior.
Cultural and personal values, in their most macro sense, give rise to differing philosophies. There may indeed be a philosophical dimension to convenience. Above and beyond apparent cultural norms, the individual's philosophy of life may affect his/her perception of convenience and whether or not he/she chooses to value convenience as a product attribute or benefit of behavior. Those 1980's consumers with a conservationist philosophy of life decry mans "convenience" products, e.g.; disposables, aerosols.
Lastly, one would expect convenience to include a situational variable. As product choice has demonstrated situational variation (Belk 1975, 1979), convenience perception and need can also be expected to vary among differing usage situations. Hornik (1982) found that time allocation was situationally influenced. As convenience admittedly includes a temporal dimension, one can deduce that in severely constrained time situations, convenience may be of greatest importance. In less time constrained situations, convenience may not be considered salient at all. Similarly, other variables of convenience may be situationally affected.
Additionally, convenience may vary depending on purchase situation. For example, convenience (in its spatial sense) may lead to situationally suggested impulse purchasing where there had not been any previous purchase intent. The inclusion of florist departments in supermarkets assumably gives rise to many impulse purchases of flowers and floral arrangements for the evening meal that otherwise would not have been considered bs the consumer.
The lack of empirical evidence supporting the simple singular time savings notion of convenience implies that the construct is more complex, with a number of variables interacting to determine the individual's need for and perception of convenience. Convenience apparently is many things to many people and it may vary among, and within, individuals along the variables just outlined.
Proposed Consumer Perceived Convenience Classes
In order to develop a more comprehensive concept of convenience, we explored a list of factors which the consumer may perceive as making a product "convenient." The many factors can be grouped into six categories, six "classes of convenience": time utilization, handiness, appropriateness, portability, accessibility, and avoidance of unpleasantness.
Table 1 lists these six classes of convenience with some product examples which we believe demonstrate each class. The six classes can be seen as marketing strategies which have been used by companies to meet consumer perceived convenience needs. Many of the products/services could be included under more than one class but each product/ service is listed only once under the class we believe is most applicable.
CONVENTIONAL CLASSES AND SAMPLE PRODUCTS/SERVICES
The time utilization class of convenience is the single type of convenience that has been operationalized in the research cited in this paper. It implies a time saving or time buying utility. But the consumer's preference for the time saving utility of a product depends on many considerations. Is there an assignable monetary opportunity cost to the consumer's use of the specific time of day, e.g., 7:00 p.m.? What amount of time does the consumer have available to allocate among consumption activities? Is the consumption time perceived as being enjoyable or creative? Can the time spent in consumption be combined with other rewarding activities, such as socializing with friends/family, watching the evening news, etc.? A most important element in choice of consumption activity may be the degree to which the activity is appreciated/rewarded by significant others, e.g., mother cooking evening meal for family.
The handiness class of convenience refers to effort saving capability. This classification represents "ease of production." A product may be viewed as convenient because it makes the consumption activity easier to perform. A food processor cuts, chops, etc. with more uniformity, speed, and with less mess than the equivalent manual activity. In addition, handiness can imply flexibility of use or production. A single appliance serves many functions, thus "easing" several different production processes.
Appropriateness refers to fittingness to specific needs. Single serving cans fulfill a specific need for single consumers. The new, under-the-cabinet appliances fill a specific need to have an appliance close at hand while still maintaining clear counter space. Gourmet frozen entrees provide singles with variety in their diet without the need to purchase and store the wide assortment of staples necessary to produce the dishes or to purchase larger than needed quantities of perishables. A further "convenience" the entrees provide is the lack of interminable leftovers that the single would otherwise accumulate.
Portability allows the consumer to consume the product in any location he/she desires. Individual packaging, dehydration, and several modern preserving methods make foods more "portable" in this convenience sense. The portable computer and the "Walkman"-type stereo allow the consumer to gain concomitant utilities, e.g., perform work while traveling and listen to music while exercising.
Accessibility is a primary class of convenience. This category includes proximity of location, availability when consumer desires product, and flexibility of delivery of product. The all-under-one-roof malls and superstores provide a variety of goods which are in the right place at the right time. Doctors and other professionals provide appointment times in the evening to be accessible when the consumer is free to partake of their services. Mail-order catalogs and videotex services allow the consumer to shop at his/her leisure at any time of day, any day of the week.
Avoidance of unpleasantness is a class of convenience in that it allows the consumer to forego an activity that he/she previously had to perform but did not enjoy. This concept includes a preference or relativity aspect. The consumer prefers other activities to activities that must be performed. The consumer would consider it"convenient" to have the necessary activities performed by someone other than him/her-self.
Figure 2 incorporates our proposed influencing variables and convenience classes into the previously illustrated convenience issues diagram (Figure 1). The proposed influencing variables increase or decrease the likelihood of a consumer choosing a convenience consumption strategy. If such a strategy is chosen the consumer will then compare products on one or more of the proposed convenience class attributes.
We are therefore suggesting that choice of convenience as a consumption strategy is a function of the previously mentioned influencing variables, and that consumer perception of a product as convenient is a function of the six just mentioned classes of attributes.
PROPOSED CONVENIENCE CONSUMPTION MODEL
IMPLICATIONS AND SUGGESTED RESEARCH DIRECTIONS
We believe it is especially important to seriously and systematically study the idea of convenience at this time because a myriad of trends foreshadow that convenience, as theoretically discussed in this paper, will become a dominant salient attribute of virtually all products consumed in American society.
Existing research has already emphasized the hypothesized effect of the increasing proportion of women working on the demand for convenience. There are also many non-time-constraining trends which increase the likelihood of heightened demand for convenience.
Americans, with increased affluence gained via attainment of high levels of education and dual income/career households, are demonstrating a new frame of mind. Housework, chores, and traditional house and family maintenance activities are being seen as menial and non-fulfilling. No longer is a woman a bad mother or unskilled woman if she can not bake a good pie or sew a practical tress. A man no longer need fear for his masculinity if he does not fix cars, mow the lawn or chop wood. The "upscale" American would simply prefer to pay others to fulfill such functions which are seen as boring or unpleasant. The affluent consumer is more and more likely to choose to devote non-working time to sports and hobbies.
The family structure of the country has changed markedly over the past 15 years, giving rise to a large minority of single person households and other non-traditional family patterns. These non-traditional households assuredly have different consumption needs and preferences than the traditional unit of husband, housewife, two kids. and a dog.
Marketers have responded to the increased level of affluence by offering products and services which could not have been afforded by a less affluent society. The mere availability of such new convenience services such as videotex and other electronic services increases the consumer's awareness of the "convenience attribute," and may increase his/her demand for "convenience" in virtually any goods.
All of these trends--increased time demands, greater affluence, increasing percentage of non-traditional households, and availability of new products--lead to the deduction that convenience as a construct and product attribute will become increasingly important to consumers, and marketers, in the future. Therefore, convenience needs much research effort in order to be utilized as a fruitful marketing construct. This concluding section represents a suggested path for future research.
First, an exploratory study is needed to gain preliminary knowledge on the perceptions consumers have concerning convenience. This exploratory effort would yield information concerning the variables influencing perceived need for convenience and information about product attributes perceived as making one product more convenient than another. Those mentioned in this paper may or may not be found to underlie convenience in the minds of consumers. Only ethnographic research can provide qualitative knowledge upon which to build the construct and lead to more valid and reliable operationalization of the construct in empirical testing.
After exploratory research has been conducted, theories and models must be generated in order to develop hypothetical links between convenience and other marketing and social science concepts and constructs. Once preliminary theories have been articulated, empirical studies need to be conducted, testing specific hypotheses. Samples representative of consumers at large need to be utilized in order to develop complete conceptualizations. Researchers need to begin to study the construct from the point of view of all household types. One can easily argue, for example, that single person households are under the severest time constraints in that there is no other household member to share in the consumption/ production process.
Finally, after empirical tests have been conducted on a variety of theoretically derived hypotheses, it will be necessary to integrate the construct of convenience into the consumer behavior model and the broad area of marketing. The construct, once satisfactorily defined, operationalized, and understood, should then be applied by marketing practitioners to more effectively and efficiently deliver the products and services needed by the modern consumer. If the aforementioned trends do indeed give rise to greater consumer need for "convenience" as a product attribute and benefit, marketers must be aware of the complexity of the construct.
Becker, G. (1965), "A Theory of the Allocation of Time, The Economic Journal, September, 493-517.
Belk, R. (1975), "Situational Variables and Consumer Behavior," Journal of Consumer Research, 2 (December), 157-164.
Belk, R. (1979), "A Free Response Approach to Developing Product-specific Consumption Situation Taxonomies," in Analytic Approaches to Product and Marketing Planning, A. Shocker, ed., Cambridge, l*A: Marketing Science Institute, 177-196.
Bucklin, L. (1963), "Retail Strategy," Journal of Marketing, 27 (January), 50-55.
Copeland, M. (1923), "Relation of Consumers' Buying Habits to Marketing Methods," Harvard Business Review, 1 (April), 282-289.
Douglas, S.P. (1976), "Cross-National Comparisons and Consumer Stereotypes: A Case Study of Working and NonWorking Wives in U.S. and France," Journal of Consumer Research, 3, 12-20.
Etgar, M. (1978), "The Household as a Production Unit," in Research in Marketing, Vol. 1, J. Sheth, ed., 79-98.
Feldman, L. and J. Hornik (1981), "The Use of Time: An Integrated Conceptual Model," Journal of Consumer Research, 7 (March), 407-419.
Hendrix, P. (1984), "Antecedents and Consequences of Time Use: Proposed Measures and Preliminary Evidence," Advances in Consumer Research, 11, 35-40.
Holton, R. (1958), "The Distinction Between Convenience, Shopping, and Specialty Goods," Journal of Marketing, 23 (July), 53-56.
Hornik, J. (1982), "Situational Effects on the Consumption of Time," Journal of Marketing, 46 (Fall), 44-55.
Hornik, J. (1984), "Subjective vs. Objective Time Measures: A Note on the Perception of Time in Consumer Behavior," Journal of Consumer Research, 11 (June), 615-618.
Jolson, M. and S. Proia (1976), "Classification of Consumer Goods - A Subjective Measure?", in Marketing: 1776-1976 and Beyond, proceedings, K. Bernhardt, ed., 71-75.
Mallan, Lucy (1968), "Financial Patterns in Households with Working Wives," unpublished Ph.D. dissertation, Department of Economics, Northwestern University.
Michael, R. and G. Becker (1973), "On the New Theory of Consumer Behavior," Swedish Journal of Economics, 75, 378-396.
Mincer, J. (1963), 'Market Prices, Opportunity Costs, and Income Effects," in Measurement in Economics: Studies in Mathematical Economics and Econometrics in Memory of Yehuda Grunfield, Stanford: Stanford University Press. 67-82.
Nicosia, F. and R. Mayer (1976), "Toward a Sociology of Consumption," Journal of Consumer Research, 3 (September), 65-75.
Reilly, M. (1982), "Working Wives and Convenience Consumption," Journal of Consumer Research, 8 (March), 407-418.
Schaninger, C. and C. Allen (1981), "Wife's Occupational Status as a Consumer Behavior Construct," Journal of Consumer Research, 8 (September), 189-196.
Strober, M. and C. Weinberg (1977), "Working Wives and Major Family Expenditures," Journal of Consumer Research, 4, 141-147.
Strober, M. (1980), "Strategies Used by Working and Nonworking Wives to Reduce Time Pressures," Journal of Consumer Research, 6 (March), 338-348.
Venkatesh, A. and N. Vitalari (1985), "Households and Technology - Some Conceptual and Theoretical Issues," in M.L. Roberts and L. Wortzel, eds., Marketing to the Changing Household, Cambridge, MA: Ballinger Publishing.
Waldran, E. and E. Jacobs (1978), "Working Wives and Family Expenditures," paper presented at the annual meeting of the American Statistical Association, San Diego, CA.