Consumer Decision Making Across Family Life Cycle Stages

ABSTRACT - This study examines the differentiation of the modernized family life cycle in the context of the consumer decision-making process. The results failed to support the hypotheses that the stages of the family life cycle could be differentiated with decision-making process variables.


Karen S. Reilly, Sevgin A. Eroglu, Karen A. Machleit, and Glenn S. Omura (1984) ,"Consumer Decision Making Across Family Life Cycle Stages", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 400-404.

Advances in Consumer Research Volume 11, 1984      Pages 400-404


Karen S. Reilly, Michigan State University

Sevgin A. Eroglu, Michigan State University

Karen A. Machleit, Michigan State University

Glenn S. Omura, Michigan State University


This study examines the differentiation of the modernized family life cycle in the context of the consumer decision-making process. The results failed to support the hypotheses that the stages of the family life cycle could be differentiated with decision-making process variables.


The family life cycle (FLC) notion is an interdisciplinary concept used in many disciplines, including sociology, psychology, economics, and marketing. It is useful as an analytical tool in studying the family (Stampfl 1978). The underlying assumption when using this tool is that individuals go through distinct evolutionary familial phases (Murphy and Staples 1979; Stampfl 1978). In marketing, particularly in the study of consumer behavior, one broad concern is the variation in consumption behavior across FLC stages (see especially, Reynolds and Wells 1977).

The purpose of this paper is to examine the relationship between stages of the FLC and the decision making process in consumption behavior. Affirmation of this relationship is important if the FLC concept is to provide substantial contribution to understanding consumer decision making. The initial debate on the discriminatory power of the FLC concept is still not resolved (see, for example, Wells and Gubar 1966; Lansing and Kish 1957; Lansing and Morgan 1955).


A variety of research approaches have been used in applications of the FLC concept. Some researchers have analyzed certain consumption behaviors across the FLC. For example, over the stages of the concept, Clark (1955) and Barton (1955) investigated the purchase decision of durable and nondurable goods, respectively, while Miller (1955) examined advertising effectiveness. Other researchers concentrated on individual stages of the FLC, such as "young singles" (Wortzel 1977), the demographic structure of "newly marrieds" (Wattenberg 1977) and the impact of children on family decision making (Ward and Wackman 1972; Berey and Polley 1968).

None of the studies in the FLC literature attempted to determine whether variables framed within a consumer decision-making (CDM) process structure varied across all FLC stages. Implied in the FLC literature is the notion that due to unique family characteristics in each stage, various consumption behaviors will be exhibited (e.g., Murphy and Staples 1979; Stampfl 1978). If this assumption is true, different decision-making characteristics would be expected at different stages of the FLC.

Since the CDM process is frequently used as a framework in consumer behavior studies, this process can serve as a useful taxonomy of behaviors to compare the different FLC stages. For example, Granbois (1963) concluded that needs vary across the FLC stages. Hence, it might be hypothesized that problem recognition would vary at different FLC stages, such as recognition of the need for certain durable goods if the decision maker is part of a family in the "young, married, with children" stage versus the "middle aged, married, no children" stage. It might be expected that the nature of passive and active search would vary as well. For example, Stampfl (1978) notes that "young, married, no children" and "young, single" groups have a higher level of shopping energy and time. This implies the hypothesis that certain decision-making characteristics such as intensity of search and the number of information sources used may vary systematically with FLC stages, with the former group hypothesized to exhibit more search activities than the latter group. It may be hypothesized also that the "married, without children." engage in more active search than "married with children" groups for similar reasons.

Additionally, the size of the evoked set and hence the nature of the evaluation phase may vary when comparing families at different FLC stages. For example, the number of considered brands in the evoked set of the "young, single" group may be greater than of the "older" group. This is suggested by the finding that age is inversely related to information processing speed and learning (Phillips and Sternthal 1977).

Further, preference for a brand may vary systematically with FLC stages. This may be inferred from varied life experiences associated with each FLC stage, including those experiences related to the product category. The product usage situation defined by the FLC stage would be expected to interact with CDM phases. For example, the "young, married, with children" group may prefer a brand of bicycles which has a family appeal, whereas the "young, married, with no children" may prefer a brand which appeals to those with a more flexible lifestyle.

Finally, it may be expected that post-purchase satisfaction will vary with the stages, perhaps due to the capabilities (or lack) of a given product to satisfy consumers who have needs structured differently by the FLC stage in which they are positioned. For example, it has been suggested that the elderly have less-formed predispositions prior to purchase than other age groups (Martin 1976). Thus, it may be hypothesized that the "older" groups may be less prone to report dissatisfaction with a purchase product than the "young" or "middle-aged" groups.

The foregoing hypotheses cover only some of the potential sets of relationships between all possible FLC stages and all possible CDM phases. Rather than test only these specific hypotheses, a broader examination of the FLC/CDM relationship is desired. If the broad examination confirms that a general systematic relationship between the FLC and CDM is present, more specific research can be advised to provide greater managerial and public policy direction.


The inferences drawn regarding the relationship between the FLC concept and the consumer decision-making process will be tested using the "modernized" FLC proposed by Murphy and Staples (1979). The modernized FLC conceptualization incorporates marital status, age, and children's ages to classify individuals and their families. As a revised version of the traditional FLC, it takes into account recent changes in family composition and life style, especially by recognizing divorce and remaining childless as modern options. There is some evidence to suggest that the modernized version is superior to the traditional form (Fritzsche 1981). The modernized FLC contains five major stages (young, single; young, married, without children; other young; middle aged: and older) with thirteen subcategories.

The following two broad hypotheses are derived:

Hypothesis 1: The five major modernized FLC stages can be differentiated with decision-making variables.

Hypothesis 2: The subcategories of the "other young" and "middle aged" major modernized FLC stages can be differentiated with decision-making variables.

The special interest in the "other young" and "middle aged" FLC groups is due to a number of reasons. It was felt that more detailed analysis of these two major groups would be more informative since a.) this is where the modernized FLC concept diverges from the more traditional FLC conceptualizations; b.) the great number of subcategories (three and six, respectively) in each major group could obscure the analysis of comparing only the major groups: c.) nearly 65% of the United States population is contained in these two major groups.




A major manufacturer of consumer durable goods collected the data for this study. [Neither the brands nor the product class is disclosed following the manufacturer's request. Published research which has examined this product class generally agrees that it is a highly involved category from the buyer's perspective. Hence. extended problem solving was expected.] Random digit dialing was used to select the subject sample, with screening to insure that the respondent was a decision maker of this product category. The sampling frame was all national area codes. A total of 2915 telephone interviews conducted by professional interviewers forms the basis for this study. Issues addressed in the questionnaire included marketing strategy development, advertising campaign tracking, promotional strategy development and consumer decision making.

The modernized FLC stages were operationalized as noted in Table 1. Due to the small number of respondents in some of the original groups, groups were logically combined to insure adequate group sizes for analysis. Five purchase decision-making phases were manipulated in the study: problem recognition, search, evaluation, preference and post-purchase. These were operationalized as follows:

Problem Recognition: This phase was operationalized by three statements relating to the respondent's need to buy in the product class. The statements were concerned with whether it was now appropriate for buying into this product class. A disagree/agree scale was used with possible values ranging from 1 to 3.

Search: This phase was operationalized in terms of passive and active search. Twenty-eight passive search variables were used, encompassing such items as recall of television, outdoor and print advertising related to the product class. Each respondent's passive search score was found by summing the number of positive (yes versus no) responses to the questions. The active search score was computed similarly, with such questions as reading product-related specialty catalogues or brochures. Twelve items were employed for this variable.

Evaluation: Evaluation was operationalized in terms of the number of brands in the respondent's evoked set. Size of the evoked set was determined by asking respondents to identify the brands that s/he would seriously consider buying if s /he were to buy the product today.

Preference: A list of five brands within the product category was read and the respondent asked to identify which s/he would most likely consider purchasing. The four possible scale values ranged from "definitely not consider" (-2) to "definitely consider" (+2).

Post-Purchase: Post-purchase feelings were measured on a five-point scale in terms of dissatisfaction (1) or satisfaction (5) with the brand currently owned.

The relationship between the foregoing twelve decision-making process variables as discriminators and the FLC stages as the criterion variables was examined through discriminant analysis. Step-wise analysis was completed with a 0.05 level of statistical significance and the Mahalanobis D2 criterion for group separation. The hypotheses were tested by examining the number and (to a much lesser extent) the nature of decision-making process variables which entered the equations with significant individual F statistics and which provided an overall significant F statistic for the discriminant models. Finally, classification or "prediction" procedures were applied as a partial test of the discriminator) efficacy of the discriminant motels. It should be noted that the objective of the hypotheses and the analyses is to discriminate among groups and not to predict. The decision-making process variables are not used to predict FLC stage membership, rather, to determine whether discrimination among the a priori FLC stages could be done.

Stages I through V of the FLC were tested first, in response to Hypothesis 1. Then, the substages within stage III and the substages within stage IV were tested corresponding to Hypothesis 2 to assess the appropriateness of the modernized FLC's expansion of these two stages.


As shown in Table 2, six of the twelve decision-making process variables were significant discriminators among the five major stages of the modernized FLC. Note that variables representing four of the five stages of CDM process entered the model, suggesting at this point of the analysis that each of the five major modernized FLC stages is unique with respect to these four CDM phases.



No decision process variables significantly discriminated among groups III (A) through (C). On the other hand, Table 3 shows the significant discriminating variables for the middle aged groups IV (A) through (D) of the modernized FLC.



Validation of the discriminant model is given in Table 4 through classification procedures. The proportional chance criterion, Cpro, was computed for each discriminant model and compared to the actual classification results. Cpro is the probability that an individual may be correctly classified by chance (Morrison 1969).

For the discriminant model involving the major FLC stages, I - V, and the model involving the middle aged subgroups, IV (A) - (D), the percentages correctly classified failed to exceed Cpro. For the discriminant model involving the young, divorced or married subgroups, III (A) - (C), there was no significant difference between Cpro and the correct classification rate.



While some decision-making process variables discriminated among the major FLC stages (Table 2) and among the middle aged subgroups (Table 3), the poor classification ability of the discriminant models diminishes the meaningfulness of the variables' "statistical significance." Hence, general interpretation of the variable means across groups and in relation to the expectations noted in the literature review may be misleading. The relatively large sample sizes may be the reason that the groups are "statistically" distinct. The results represent a particularly poor showing since the sample used to derive the discriminant models was the same sample employed in the validation procedure. Ordinarily, use of identical samples should lead to a biased high correct classification result.

Given the lack of support for the hypotheses, an attempt was made to refine the analysis, post hoc. It was suspected that the findings which would lead to the support of the hypotheses might be obscured by the failure to account for differences in husband and wife decision making for those respondents who were married. Studies have shown differences between husband and wife decision making in the family purchasing context (e.g., Davis 1976; Curry and Menasco 1979; Filiatrault and Ritchie 1980). Consequently, an attemPt was made to control partially for this.

The data would not allow the direct control of both the husband and the wife in a given family since only one of the two in a married household was asked to participate in the study. However, gender information was collected as a collateral item during the interviews. Gender was then used as a surrogate for husband/wife status. FLC groups composed of married households (58% of the total sample) were re-analyzed with gender controlled.

Two sets of discriminant analyses were run. In both sets, gender was forced as the first variable in the step-wise procedure. The first set used the young, married subgroups [III (B)] as the criterion variables, while the second set of analysis used the middle aged, married subgroups [IV (A), IV (C)]. In the first set, none of the decision-making process variables was significant as a discriminator, which was consistent with the original uncontrolled analysis. In the second set, the same two variables as in the original analysis shown in Table 3 were significant but with little improvement in classification accuracy.


Little support was found for the hypotheses. Discrimination among the major stages of the FLC through consumer decision-making process variables could not be validated through classification procedures. Further, the g-eater specificity of the modernized FLC does not appear to enhance differentiation among the stages in the context of consumer decision making. These findings are particularly unfortunate since the use of a high involvement product class would lead one to expect strong positive findings. Additionally, the attempt to control for a potential husband/wife and FLC interaction did not provide support for the hypotheses.

There are several reasons why these results may have occurred. Many researchers (see Stampfl 1978, for a review) have suggested that different types of purchases (focusing on the purchase decision-making phase) are related to different stages of the FLC. Perhaps the FLC concept is useful for that general application, but not for the other phases of the consumer decision-making process.

Another possible explanation of the negative findings is related to the questioning context. Each respondent was asked to recall past behavior or predict future behavior to measure all phases of the decision-making process. It may be more suitable to match respondents by CDM phase and FLC stage prior to examining the relationship between the variables associated with a given phase and the FLC stages. The CDM phase the respondent happens to be in when interviewed may systematically bias his or her responses to questions related to previous or subsequent CDM phases.

The foregoing issue may also be related to recency of purchase, uncontrolled in this analysis. Although it would be expected that CDM differences among FLC stages would be observable regardless of purchase recency, the "true" FLC/ CDM relationship may have been obscured in the present analysis if purchase recency is positively correlated with recall ability.


Since other studies (e.g., Clark 1955: Barton 1955: Fritzsche 1981) found the FLC to be useful, it is somewhat unsettling to conclude that it has no discriminatory power with respect to consumer decision making phases. Further research can be suggested to determine whether the relationship between CDM and the FLC does in fact exist.

First, the variables which comprise the index for FLC need closer scrutiny. Perhaps there are better cut-off points for age of household heads and numbers and children. Perhaps marital status needs to be measured in terms of length of the marriage as opposed to simply a categorical variable.

Second, research is needed to determine the scale "value" for each FLC stage. The optimal combinations of FLC characteristics (that is, age, marital status, absence or presence of children within different age ranges, and so forth) need to be empirically determined for each stage.

Similarly, and third, the CDM variables and their combinations should also be determined to form better measures of the phases of consumer decision-making. For decades, consumer researchers have developed and applied scales of various theoretical concepts designed to measure portions of the domain of each and every CDM phase. Despite constant reference to the overall decision-making process as the basic behavioral paradigm, the discipline has yet to develop properly designed instruments to measure the phases of this basic paradigm.

Finally, with these refined measures, the present study can be replicated over a variety of product categories. The applicability of the FLC in combination with CDM should be determined by varying characteristics of the product categories, such as durable versus non-durable, high versus low involved. and roods versus services.


Berey, L.A. and R.W. Pollay (1968), "The Influencing of the Child in FamilY Decision Making," Journal of Marketing Research, 5, 70-72.

Blalock, H.M. (1979), Social Statistics, New York: McGraw Hill, 602.

Clark, L.H. (ed.) (1955), Consumer Behavior, Vol. 2, New York: New York University Press.

Curry, D.J. and M.B. Menasco (1979), "Some Effects of Differing Information Processing Strategies on Husband-Wife Joint Decisions," Journal of Consumer Research, 6, 192-203.

Davis, H.L. (1976), "Decision Making Within the household," Journal of Consumer Research, 2, 241-260.

Duvall, E.M. (1971), Family Development, 4th ed., Philadelphia: J.B. Lippincott Co., 106-132.

Filiatrault, P. and B. Ritchie (1980), "Joint Purchasing Decisions: A Comparison of Influence Structure in Family and Couple Decision-Making Units, " Journal of Consumer Research, 7, 131-140.

Fritzsche, D.J. (1981), "An Analysis of Energy Consumption Patterns by Stage of Family Life Cycle," Journal of Marketing Research, 18, 227-232.

Granbois, D.H. (1963), "The Role of Communication in the Family Decision Making Process," in Proceedings, ed. S. Greyser, Chicago: AMA, 44-57.

Lansing, J.B. and L. Kish (1957), "Family Life Cycle as an Independent Variable," American Sociological Review, 22, 512-519.

Lansing, J.B. and J.N. Morgan (1955), "Consumer Finances Over the Life Cycle," in Consumer Behavior, Vol. 2, ed. L.H. Clark, New York: New York University Press, 36-51.

Martin, C. (1976), "A Transgenerational Comparison: The Elderly Fashion Consumer, " in The Elderly Consumer, ed . F.E. Waddell, Columbia, MD: The Human Ecology Center, Antioch College, 251-259.

Miller, D.K. (1955), "The Life Cycle and The Impact of Advertising," in Consumer Behavior, Vol. 2, ed. Lincoln H. Clark, New York: New York University Press.

Morrison, D.G. (1969), "On the Interpretation of Discriminant Analysis, " Journal of Marketing Research, 6, 156-163.

Murphy, P.E. and W.A. Staples (1979), "A Modernized Family Life Cycle," Journal of Consumer Research, 6, 12-22.

Phillips, L.W. and R. Sternthal (1977), "Age Differences in Information Processing: A Perspective on the Aged Consumer, " Journal of Marketing Research, 14, 444-457.

Reynolds, F.D. and W.D. Wells (1977), Consumer Behavior, New York: McGraw-Hill

Rodgers, R.H. (1973), "The Family Life Cycle Concept-Past, Present and Future, " paper presented at 13th International Family Research Seminar, Committee on Family Research International Sociological Association, Paris, France.

Stampfl, Ronald W. (1978), "The Consumer Life Cycle," Journal of Consumer Affairs, 12, 209-217.

Ward, D. and D.B. Wackman (1972), "Children's Purchase Influence Attempts and Parental Yielding, " Journal of Marketing Research, 9, 316-319.

Wattenberg, B.J. (1975), "The Forming Families: The Spark in the Tinder, 1975-1985," in 1974 Combined Proceedings, ed. R.C. Curhan, Chicago: AMA, 51-62.

Wells, W.C. and G. Gubar (1966), "Life Cycle Concept in Marketing Research," Journal of Marketing Research, 3, 355-363.

Wortzel, L.H. (1977), "Young Adults: Single People and Single Person Households," in Advances in Consumer Research Vol. 4, ed. W.D. Perreault, Jr., Atlanta: Association for Consumer Research, 321-329.



Karen S. Reilly, Michigan State University
Sevgin A. Eroglu, Michigan State University
Karen A. Machleit, Michigan State University
Glenn S. Omura, Michigan State University


NA - Advances in Consumer Research Volume 11 | 1984

Share Proceeding

Featured papers

See More


Dancing with Commercialism: Emphasizing Dramatism to Persuade

Yuxin Bai, Lancaster University, UK
Xin Zhao, Lancaster University, UK
Hayley Cocker, Lancaster University, UK

Read More


Dimming the Light Offers A Creative Lens: The Impact of Ambient Illuminance on Creativity Assessment

Chen Wang, Drexel University, USA
Ravi Mehta, University of Illinois at Urbana-Champaign, USA
Rui (Juliet) Zhu, Cheung Kong Graduate School of Business, China
Jennifer Argo, University of Alberta, Canada

Read More


Better Marketing for a Better World

Jonah Berger, University of Pennsylvania, USA
Jonathan Zev Berman, London Business School, UK
Darren Dahl, University of British Columbia, Canada
Markus Giesler, York University, Canada
Rebecca Hamilton, Georgetown University, USA
Gita Venkataramani Johar, Columbia University, USA
John Lynch, University of Colorado, USA
Andrea Morales, Arizona State University, 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.