Determining the Durable Product Needs of Households


John F. Willenborg (1971) ,"Determining the Durable Product Needs of Households", in SV - Proceedings of the Second Annual Conference of the Association for Consumer Research, eds. David M. Gardner, College Park, MD : Association for Consumer Research, Pages: 108-118.

Proceedings of the Second Annual Conference of the Association for Consumer Research, 1971     Pages 108-118


John F. Willenborg, University of South Carolina [Assistant Professor of Marketing, University of South Carolina.]

The primary purchasing unit of society is the household. Purchases of durable goods, as well as many nondurables, are made principally for household consumption. Even when an actual purchase is made by one family member, the household's influence is likely to be substantial, particularly for durables. For example, the purchase of a stereo console with mahogany sliding doors, priced at $445, and featuring a Garrard record changer, may represent the combined preferences of several different household members. In fact, the stereo itself may have been purchased as a compromise between a color television set and a radio-TV-stereo combination desired by two different influential household members.

Despite such rather obvious influences, buyer behavior research generally concentrates on individuals and not on the households of which they are a part. There are some very pragmatic reasons for this approach, including the difficulty of measuring interaction between household members and the cost of obtaining suitable data.

This researcher maintains, however, that durable product needs of households can be determined via longitudinal analysis. Thus, the focus in this study is on a method for identifying the durable product needs of households and relating them to subsequent purchasing and product-related behavior. [Durables are chosen for study because the likelihood of measurable household influence is greater than for nondurables.]

In this study, a central role is assigned to attitudes toward product class characteristics. It is not implied that other variables are insignificant factors in decision-making, but rather that strong positive relationships can be found between these attitudes and behavior for individual households and that such relationships are valuable indicators of future behavior. Therefore, potential buyers' attitudes toward certain attributes of durable goods are measured to obtain an index of the likelihood of purchase.

It is contended that the image or concept of a product as viewed by the household members is of extreme importance to purchasing behavior. Substantial support for this notion is found in comprehensive decision models formulated in recent years which consider the influence of product characteristics. For example, the Nicosia decision process model includes as variables the technical, functional, and social psychological attributes of a product or brand which may or may not lead to a purchase. [Francesco M. Nicosia, Consumer Decision Processes: Marketing and Advertising Implications (Prentice-Hall, Inc., Englewood Cliffs, New Jersey, 1966), pp. 157-58.]

An interesting, related view is that held by Walter A. Woods, who distinguished between two sets of variables in choice behavior--consumer and product variables. Consumer variables refer to individual differences in cognitive structures while the latter have to do with the character of the product itself. Woods maintains that certain classes of products generate demand for different reasons, for example, degree of ego-involvement generated, hedonistic qualities, and status. [Walter A. Woods, "Psychological Dimensions of Consumer Decision," Journal of Marketing, Vol. 24 (January, 1960), p. 16.] The importance of the Woods thesis to this study lies in the recognition that the characteristics of products themselves may have an influential role in behavior.

Several other studies of specific product and brand dimensions are noteworthy. For example, articles by Yankelovich, [Daniel Yankelovich, "New Criteria for Market Segmentation," Harvard Business Review, Vol. 42 (March-April, 1964), pp. 83-90.] Kuehn and Days [Alfred A. Kuehn and Ralph L. Day, "Strategy of Product Quality," Harvard Business Review, Vol. 40 (November-December, 1962), pp. 100-110.] and McClure and Ryans [Peter J. McClure and John K. Ryans, Jr., "Differences Between Retailers' and Consumers' Perceptions," Journal of Marketing Research, Vol. 5 (February, 1968), pp. 35-40.] report on studies of product attributes. Recently, Lehman [Donald R. Lehman, "Television Show Preference Application of a Choice Model," Journal of Marketing Research, Vol. 8 (February. 1971). pp. 47-55.] has utilized a preference model based on such attributes.


The research upon which this paper is based assumes a positive relationship between attitudes of household members toward durable goods attributes and purchasing behavior. The ultimate goal of most attitude studies is, of course, prediction of behavior. Obviously, this behavior is seldom completely predictable. However, it is possible for the marketer to ascertain "tendencies" of potential consumers so that he may act either to reverse unfavorable ones or to capitalize on apparent opportunities. Thus, after determining which households exhibit need for a product or are "in the market" for one, he can plan appropriate action. It is important to recognize that product needs are present in varying degrees. A household, in effect, will establish either formally or informally a hierarchy or priority system for the purchase of products. [John McFall, "Priority Patterns and Consumer Behavior," Journal of Marketing, Vol. 33 (October, 1969), pp. 50-55. McFall maintains that sets of durables are planned and ranked for purchase planning.] Underlying attitudes toward characteristics of the products are, at least, partially responsible for the hierarchy formation. Thus, it is logical to attempt to measure such attitudes in order to determine the hierarchical listing of durable goods priorities; i.e., the household "need structure."

The Household Durable Product Need Structure

It is postulated that the higher a product ranks in a household's product need structure, the more likely it will be purchased before a lower-ranked product. Similarly, "product-related-behavior" will tend to be undertaken to a greater degree for high priority products than for low priority ones. Thus, the research hypothesis is stated:

The household need structure of durable goods priorities, as established through the measurement of attitudes, is an effective indicator of subsequent purchasing and product-related behavior over time.

The bases used in the study for the development of the household product need structure are the expressed attitudes of potential consumers toward attributes of products. In developing the need structure, it is assumed that the consumer's overall attitude is composed of several elements. As determined in a pre-test, the general attributes of the products which are considered to be components of attitude are the product's necessity, affordability, efficiency, and attractiveness. (See Exhibit 1)



The consumer may be favorable toward some and unfavorable toward other product attributes. In addition, varying degrees of favorability or unfavorability may be present. Obviously, for a purchase to be made or for other product-related behavior to be undertaken, the sum of one's attitude components-however he weights their relative significance--will be positive, resulting in an overall attitude which is favorable to some degree.

Product-Related Behavior

The research assumes the functional relationship:      PT = f(Se, O, I, C, So, B, Own)

That is, the tendency of a household to purchase (PT) a durable good is a function of the incidence and level of product-related activity undertaken by members of the household. The activities include search (Se), observation (O), inquiry (I), consultation (C), social interchange (So), brand recognition (B), and product ownership (Own) by others (Exhibit 2). These elements are incorporated into behavior indices for comparison with valuations of products as per the need structure. Implicit in the procedure is the assumption that, because a purchase often culminates from such activity, it is considered to be desirable behavior from a marketer's standpoint.



Research Design

Members of a panel of fifteen households (generally classified as "young marrieds") were interviewed in five waves over a period of approximately six months. In three of the interviews (#1, #3, and #5), need structures of the household were determined by measuring attitudes toward durable product attributes (necessity, affordability, efficiency, attractiveness). In each case, husband and wife rated the attributes of specific durables independently as well as through a joint decision-making process. The list of durables subject to analysis and subsequently making up each household's need structure was composed of twenty or more products commonly purchased for household use. [The basic list of durable goods is shown in Exhibit 3. Certain other items were included for individual households based on interests and intentions expressed by household members.]

The attitude measurement technique of magnitude estimation [Magnitude estimation is a subjective ratio measurement developed in psycho-physics, largely by Stanley S. Stevens. Experiments have shown that there exists a recognizable lawful relation between physical stimuli and the perception of them. The magnitude of a sensation increases as a power function of the magnitude of the stimulus as in the form Y = koB   where Y is the perceived magnitude, k is a constant, o is the physical value, and B is the exponent of the power function. Magnitude estimation has been applied in numerous experimental situations involving nonphysical stimuli. Stevens has commented: ". . . magnitude estimation. . . has been used to gauge the consensus concerning intensity or degree for such variables as strength of expressed attitudes, pleasantness of musical selections, seriousness of crimes, and other subjective dimensions for which the stimuli can be arrayed only on nonmetric or nominal scales" (in Stanley S. Stevens, "A Metric for the Social Consensus," Science, Vol. 151, 1966, p. 530). See also: Robert L. Hamblin and Carole R. Smith, "Values, Status, and Professors," Sociometry, Vol. 29, 1966, pp. 183-196; and Robert L. Hamblin, "Ratio Measurement and Sociological Theory: A Critical Analysis," (St. Louis, Missouri, Washington University, mimeograph), 1966, p. 17.] was utilized to first determine the relative importance to respondents of each product attribute in the purchase of various classes of durables. For this purpose, durables were placed in one of the following categories: household furnishings, kitchen appliances, cleaning products, and entertainment products.

Household members were then requested, via magnitude estimation, to provide a measure of the degree of necessity, affordability, efficiency, and attractiveness which they ascribed to each durable in the listing. The estimates were weighted according to the relative importance of each attribute as determined earlier and the resultant values for each product summed. The summations were ranked in descending order to provide hierarchies of product preferences (need structure) for individuals and for the household (Exhibit 4).



In subsequent interviews, need structures were again determined. These were compared, in order to test the research hypothesis, with an index of behavior reflecting the product-related activities and ultimate product purchases by respondents. [Three indices of behavior were calculated using different weighting procedures. No significant difference was found between need-structure behavior correlations using each of the three indices; therefore, only one index is used in the analysis and is explained in Exhibit 5.] The behavior was determined by direct questioning in interviews #2. #4. and #5.




The hypothesis, suggesting a strong relationship between need structure ranking and product-related behavior, was verified by the data generated in the study. The testing of the hypothesis was carried out in three ways:

1. By comparison of the levels of activity associated with products ranked in various segments of the need structure.

2. By calculation of correlation coefficients measuring the relationships between product values and behavior indices.

3. By determination of the relationship between product need structure position and actual product purchases.

It should be noted that although the household need structures include both owned and unowned products, products not presently owned by households are of greater concern because of their likelihood of purchase by young married couples. Owned products are of interest only where replacements are in order or where households wish to own more than one of a certain durable. Thus, for purposes of analysis, presently unowned products receive primary consideration



Product-Related Behavior and Need Structure Segments

For the first test of the hypothesis, the need structure ranking--as determined in the first interview with each family--was arbitrarily divided into four "quartiles." These segments were selected simply to provide a convenient way of showing declining levels of behavior associated with products with successively lower need structure rankings. Arithmetic means of the behavior indices were computed for presently unowned products in each quartile of the total need structure. In nearly every case, the means were lower for each lower segment beginning with the top quartile. (See Exhibit 6)



A related test was to compare the number of products associated with behavior to the number with no behavior. Of 150 products classified as unowned, between 81% and 90% ranking in the highest quartile were associated with some activity, between 74% and 88% in the second quartile, between 42% and 48% in the third, and only between 18% and 20% in the fourth were related to behavior of any kind when husband, wife, and joint rankings were considered

Correlations Between Need Structure and Behavior

Pearson product-moment correlation coefficients were computed to test the relationship between household product need structures and product-related behavior. The magnitude estimates provided by husband and wife jointly, as well as independently, were correlated with the measures of behavior associated with each product.

Adopting the conventional .05 level of significance (based on t-test), the correlations were found to be significant for nine of the fifteen households. In five cases, correlations were significant at the .01 level. Exhibit 7 shows the correlations based on joint household rankings.

The correlation levels varied over a fairly wide range. However, in only three of the fifteen households was the correlation coefficient lower than .50 for all three need structures (husband, wife, and joint). In five cases, correlations for at least one household member were higher than .70. Given the complexity of the components of the product values and the arbitrary weighting of behavior, the correlations were judged to be rather remarkable. An example is provided in Exhibit 8 of the da,a from one household, which was rather active regarding products, When judged in relation to levels of activity associated with the various products, the .7295 correlation coefficient is seen to be very good for a study of this nature.

Need Structure and Actual Purchases

Over the period of the study, the households acquired forty of the products which had been listed and ranked in their respective need structures. Analysis of the data showed that, with the exception of three acquisitions (two gifts and the other classified by the buyer as "unnecessary" and "on impulse"), each product had been ranked by at least one household member among the upper seven products in the household's (including both owned and unowned durables) need structure in the interview immediately prior to its acquisition. In addition, of the thirty-seven products, only three were ranked as low as tenth in the need structure by the husband and wife evaluating products jointly.





When only unowned products were considered, no purchased products ranked lower than fifth among unowned products when ranked by husband and wife together. In addition, the following observations were made: (1) thirty-three products, when acquired, were ranked either first or second by either husband or wife or both; (2) seventeen products were ranked first by husband and wife acting together; (3) of the seven gifts, five had been ranked either first or second by at least one household member prior to acquisition by the household.

In a follow-up study six months after the original data were collected, it was found that fifteen more products listed had been purchased in the interim. As might be expected, given the longer intervening time period, need structure rankings for some purchases were not as high. However, thirteen were ranked in the upper 50% of both owned and unowned products in the last interview prior to purchase by husband and wife together. Of the fifteen purchases, 12 were previously unowned by the households, of which only three were ranked lower than fifth among unowned products. It may be speculated that strong favorable attitudes toward durables will ultimately result in purchases and that these attitudes are slow to change over time. Seemingly, the results support the McFall thesis that households desire a complete set of durables and will continue to rank each durable highly until purchase is made



The general objective of the research--to explore the feasibility of using a measure of consumer household attitudes in order to indicate the extent of subsequent purchasing and other behavior related to durable goods--was clearly fulfilled. Results of tests of the hypothesis are summarized below.

The products in the need structure were related, in various ways, to subsequent purchases and levels of product-related behavior. A direct relationship was found between need structure ranking and extent of behavior. A considerably higher level of product-related activity was associated with higher-ranked products than with lower-ranked ones. Also, purchases tended to be related to need structure position. Products generally were not purchased unless consumer attitudes toward them were strongly positive, as reflected in high product valuations. Even gifts and "impulse" purchases tended to be correlated with the need structure. Therefore, the hypothesis was given strong support.


The research results lend themselves to the suggestion of methodological adaptations and research directions, some of which are listed below.

1. Incorporation of other product attributes such as durability and dependability into the analysis could have an interesting effect on the product valuations and, thus, upon the product preference listings.

2. The variable of the relative influence of husband and wife on purchasing decisions could be incorporated into the methodology.

3. The analysis could be extended to measuring attitudes toward specific brands within product classes.

4. Research efforts could be directed toward the determination of whether consumers actually consider products in "categories" such as entertainment, cleaning, and household furnishings and, if so, how heavily they weight the attributes of products within each classification.

5. The question of which, if any, product attribute is the dominant determinant of purchase could be investigated further. Alpert [Mark I. Alpert, "Identification of Determinant Attributes: A Comparison of Methods," Journal of Marketing Research, Vol. 8 (May, 1971), pp. 184-191.] has suggested some methodological considerations. This research has relied on the ability and willingness of respondents to estimate relative importance.

6. The study could be extended to established households in which a majority of purchases would be of a replacement nature.

7. If a strong, favorable attitude toward a product does indicate a "tendency" to undertake action directed at purchase, as this research indicates, it would be a natural outgrowth of this study to subject households to varying degrees of promotional effort relating to high and low-ranked products to test the strength of the inclination to purchase.



John F. Willenborg, University of South Carolina [Assistant Professor of Marketing, University of South Carolina.]


SV - Proceedings of the Second Annual Conference of the Association for Consumer Research | 1971

Share Proceeding

Featured papers

See More


The psychological impact of annuities: Can pension payout choice influence health behavior?

Anja Schanbacher, London Business School, UK
David Faro, London Business School, UK
Simona Botti, London Business School, UK
Shlomo Benartzi, University of California Los Angeles, USA

Read More


Morality Matters in the Marketplace: The Influence of Morally Based Attitudes on Consumer Purchase Intentions

Andrew Luttrell, Ball State University
Jacob Teeny, Ohio State University, USA
Richard Petty, Ohio State University, USA

Read More


How Numeric Roundness Influences Probability Perceptions

Julio Sevilla, University of Georgia, USA
Rajesh Bagchi, Virginia Tech, USA

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.