Measuring Family Decision Making: Problems and Prospects

Michael P. Heffring, University of Pittsburgh
ABSTRACT - The family as a unit of analysis represents a common sense but complex way to view consumer decision making. This perspective involves viewing the family both as a statistical group as well as a primary social group. The latter view underscores a concern not only for identifying the diversity in family role structures at an aggregate level but focusing on the processes of communication, interaction and influence that occur within specific family role structures. To do so requires increased attention co the definition of roles; specification of decisions as they relate to roles and at what level; using roles to determine who to talk to; viewing response incongruency as a function of role-taking inaccuracy and an attitude behavior discrepancy; and integrating both self-report and observational methods.
[ to cite ]:
Michael P. Heffring (1980) ,"Measuring Family Decision Making: Problems and Prospects", in NA - Advances in Consumer Research Volume 07, eds. Jerry C. Olson, Ann Abor, MI : Association for Consumer Research, Pages: 492-498.

Advances in Consumer Research Volume 7, 1980     Pages 492-498

MEASURING FAMILY DECISION MAKING: PROBLEMS AND PROSPECTS

Michael P. Heffring, University of Pittsburgh

ABSTRACT -

The family as a unit of analysis represents a common sense but complex way to view consumer decision making. This perspective involves viewing the family both as a statistical group as well as a primary social group. The latter view underscores a concern not only for identifying the diversity in family role structures at an aggregate level but focusing on the processes of communication, interaction and influence that occur within specific family role structures. To do so requires increased attention co the definition of roles; specification of decisions as they relate to roles and at what level; using roles to determine who to talk to; viewing response incongruency as a function of role-taking inaccuracy and an attitude behavior discrepancy; and integrating both self-report and observational methods.

In an extensive review article by Davis (1976), three major areas of research on family decision making were identified:

- involvement of family members in economic decisions

- the process by which family decisions are made

- the consequences of different family structures and decision making styles.

Considerable headway has been made in the study of involvement (Blood-Wolfe, 1960; Cunningham and Green, 1974 Wolgast, 1958; Ferber and Lee, 1974; Davis and Rigaux, 1974; Scanzoni, 1977; Rigaux, 1978), and some early innovative efforts are being made in studies of the process (Weik, 1971; Hempel and Ayal, 1977; Webb, 1978) as well as consequences (Hunt, 1977; Heffring, 1978).

Although this increased effort in studying family decision making has led to a number of interesting conceptual and methodological findings, it has also led to a number of interesting conceptual and methodological problems. The word interesting is used Co describe these problems because it is through the systematic investigation of interesting problems (or anomalies as Kuhn (1962) puts it) that consumer researchers are likely to wake significant conceptual and methodological gains. Lack of recognition or avoidance of these problems no doubt leads to many of the deficiencies Jacoby (1978) has indicated are rampant in consumer behavior research today.

What are some of the problems in family decision making research? They include the following:

(i) the family as a unit of analysis: statistical a social group?

(ii) family role structures: what are we measuring?

(iii) self-report vs. observational measures: problem or prospect?

The Family As a Unit or Analysis : What Does It Mean?

The term "family" elicits a perplexing problem in research; it is a term with almost universal recognition and emotional appeal. At one time or another, researchers and consumers alike have experienced being part of a family of one or another. However, this widespread recognition is matched by an equally widespread diversity in the form, function and meanings of the different families we have experienced. Thus, while communication is facilitated by recognition; understanding and common use of the term is marred by diversity in meanings.

On a denotative level there is an interchange of the terms household, family, nuclear family, conjugal family and extended family. At the broadest level, a household consists of all persons who occupy a housing unit. [These definitions follow from those used by the Bureau of Census, Current Population Reports, Series P-20, No. 324, "Population Profile in the United States," U.S. Government Printing Office, Washington, DC, 1978.] This includes people living alone, a group of unrelated people or a family (as defined in the census as a group of two or more persons related by blood, marriage, or adoption and residing together). Note that these definitions are derived from an analysis of social living arrangement at a particular point in time.

In 1977, there were 74.1 million households. [Ibid. p. 20.] Of these, 56.5 million constituted families. These families all represent nuclear families but not necessarily conjugal families. A nuclear family refers to two or more persons who are related to one another by blood, marriage or adoption and are of the same or adjoining generations (e.g., one parent and a child, husband-wife and children, brother and sister, etc.). The term conjugal, however, implies marriage or the relation between husband and wife as the basis of the structure. Extended families refer to family structures that extend beyond the nuclear family - e.g., multiple numbers of nuclear family groupings.

Further breakdowns of the above mentioned categories leads to even finer definitions of what can constitute a family:

- 15.9% of all households include a father as the sole wage earner, a mother as a full time homemaker and at least one child.

- 18.5% include both the father and the mother as wage earners, plus one or more children at home.

- 30.5% of households consist of married couples with no children, or with no children living at home.

- 6.2% are headed by women who are single parents, with one or more children living at home.

- 0.6% are headed by single parent males, with one or more children at home.

- 2.5% consist of unrelated people living together, (including couples co-habitating).

- 20.6% are single person households.

- 5.3% consist of female or male headed households that include relatives other than spouses.

Thus, consumer researchers are faced with a multitude of different demographic descriptions of families from which to sample.

It is these types of demographic or statistical representations that are primarily used when the household or family is referenced to in the consumer literature. Families are described in terms such as life cycle, household size, household income, occupational status, and household composition as well as other demographic variables. Granbois, Summer and Frazier (1977) use some of these variables in assessing normative expectations, predictive expectations, and problem incidence. Wall, et al. (1977), Miller (1977) and Mason and Hines (1973) have also used these variables. At the level of prediction these variables may be reasonable representations of aggregate household behavior. However, they tell you little about the processes of communication/ interaction and influence that are occurring within particular households and families, or why households having similar demographic profiles may be using totally different processes to make and subsequently evaluate, consumer related decisions. A/though many studies compare aggregate consumption behavior, relatively few studies have focused on the different interaction and influence patterns that exist between families where the wife works or doesn't; where there are children present or not; whether there is one parent or two; and whether the couples are married or not. It is clear that not only are the processes different but also what the word "family" means to these different groups varies tremendously.

To analyze and research these processes and "meanings", an alternative approach is to focus on families (as the predominant household type) and view them not only as STATISTICAL GROUPS (that is, groups formed not by members themselves but by statisticians, academic researchers in consumer behavior, or marketing researchers) but also as SOCIAL GROUPS (groups in which the members actually interact and associate with one another; each member conscious of a sense of his belonging to the group). In a model building perspective, variables associated with this added dimension may more completely specify the model and greatly improve its explanatory as well as predictive power. To the extent that variables representing this dimension correlate with variables (independent and dependent) already in the model, it is merely a matter of making these relationships explicit. This approach is consistent with the recent work of Rigaux (1978).

The Family As a Primary Social Group

To extend the previous point, it is useful to view the family as a primary social group, as characterized by the following:

* face to face contact

* smallness of size

* frequent and intense contact

It differs from most ad hoc primary groups in that there are prescribed roles, a sense of permanence and a history as characterized by recurrent decision situations.

Of the variety of perspectives and conceptual frameworks that have been used to study the family (structural-functionalism, situational-determinism, institutionalism, developmentalism and interactionism, Hill and Hansen (1960), Broderick (1971)); it is the interactionism perspective and its variant - "social interactionism" (Bagozzi and Loo, 1978) that most clearly embodies the concept of the family as a primary social group. From this perspective, the family is represented as a system of dyads.

Social exchanges occur within these dyads, the outcomes of which are a function of the shared, socially constructed attitudes of family members as mediated by the social processes of power, conflict, decision-making and influence. [Bagozzi, R. P., and M. F. VanLoo, "Fertility as Consumption: Theories from the Behavioral Sciences," Journal of Consumer Research, (March, 1978) p. 224.] It is assumed that the contents of these exchanges have both utilitarian as well as symbolic meanings (e.g., a car is not only a mode of transportation but a source of potential status and prestige). These meanings are social products derived from, and consistently modified through, the interactions and exchanges that occur between family members.

Further, it is assumed that certain forces constrain exchange relationships. These may occur from either outside the exchange (e.g., conditions in the economy, cultural norms, social circumstances, etc.) or within (e.g. individual needs, motivations, and personalities as well as patterns of power, conflict and cooperation that have been established.)

As applied to the area of consumer behavior, this perspective implies that a number of new concepts should be introduced into studies of family decision making and certain modifications made to concepts currently being used. One area in particular where both the former and latter can be applied is in the area of family role structures.

FAMILY ROLE STRUCTURES: WHAT ARE WE MEASURING?

Although much attention has been paid to the involvement of family members and the roles they play in decision-making, there has not been a consistent definition or use of the term role in the family decision making literature. Like the term "family" there is widespread recognition matched by diversity in use. Roles are institutionalized expectations about behavior associated with particular social position or statuses. Nadel (1957) identifies four necessary conditions for labeling a behavior pattern a role [NOTE: For an extended discussion of how the role concept can be applied to consumer research in general see Wallendorf, M , "Social Roles in Marketing Contexts," American Behavioral Scientist, 21 (March/April, 1978) 571-82.]:

(i) Behavior: roles refer to people's behavior not strictly their attitudes or intentions.

(ii) Interaction: roles refer to relationships between two or more people. One is a husband because of the interaction (actual, potential, or previous) with a wife (and vice-versa).

(iii) Rules: roles are rule based. They are derived from expectations individuals have concerning people's behavior who hold particular positions. These may be cultural expectations concerning what constitutes a good parent or a good mother; they may also be specific expectations that have arisen from interactions in a particular family (i.e. family history).

(iv) Bundle of behaviors: roles indicate a number of behaviors or functions that the incumbent is supposed to perform and fulfill. For example, the husband role may include the following: providing monetary support to keep him and his spouse sustained, being "faithful" to his spouse; and providing emotional support to his spouse (e.g. affection, companionship, etc.). It should be stressed that there may not be total consensus on these expectations: execution of these expectations may vary by culture and/or family type. However, despite the differences in execution there is still agreement on the behaviors expected of husbands in general.

What are the implications of this definition of role for family decision making research?

(i) Increased specification of role conditions

It seems that the majority of research has tapped the rule/expectation based dimension of roles with lesser amounts of work done or specifying what set of behaviors are encompassed in the role (the bundle), the interaction that results from these expectations, or the subsequent behavior and the degree of satisfaction associated with it. Davis (1976) cites that the agreement by couples on roles played by family members varies from 30 to 80 percent. Although Davis cited methodological reasons for the variances an equally compelling argument can be mede that the variance is a result of measuring only on-.-e-e necessary condition of a role while three other necessary conditions are not measured or are measured less precisely.

(ii) Role sampling

The research to date generally assumes that what is being measured is the relative influence of the husband and wife. This assumes that these roles individuals are playing are in fact the HUSBAND role and the WIFE role. Again, a strong argument can be made that what in fact is being measured in these studies is a wide variety of different roles: husband-wife, mother-father, executive-housewife, boyfriend-girlfriend, bridegroom-bride, teacher-learner, etc. Since involvement and consensus over involvement has been shown to vary by product category, it is posited that this variance is tied to the execution of different roles with different products. For example, the mother-father roles are likely to be more dominant in school and children's toys decisions, while the husband-wife roles dominate in car and saving objectives decisions. This introduces a new area of sampling in family decision making research. That is, researchers must be cognizant of the fact that of all the different roles that individuals in a family might play, they are researching only a sample of these. Those roles chosen for research should be specified so that the roles not researched are also clearly stated and recognized. This would help reduce possible specification errors in the construction of family decision making models. Role sampling would also take into account the fact that certain roles become more important as a family moves through the life cycle - the growing importance or decline of particular role being positively correlated with the importance of particular product decisions. Currently, researchers construct influence scales that weight the importance of various product decisions equally. Price-Bonham (1976) has tried different weight estimation procedures but results have been indeterminate. Better weight deter-ruination procedures might be developed using a role sampling procedure.

(iii) Decision at what levels

Use of the role concept also sensitizes the researcher to the possibility that decisions may not be made at the same level. Figure 1 identifies decisions at differing levels of abstraction. Due to the roles played by family members, it may be that one member is responsible for decisions at the higher levels while the other member is responsible for decisions at the lower levels. For example, given current measurement procedures, a wife may appear dominant a generic product area (e.g. food purchases), but this may be a function of delegation rat-her than influence or power. That is, based on the prevailing role pattern in the family, the husband may be responsible for general economic matters (similar to Ferber and Lee's (1974) Family Financial Officer), and in fact delegates responsibility for particular decisions to various family members. This may be a function of such traditional sources as expertise, as well as authority (the prescribed right to make decisions based on cultural or social expectations/norms).

In a "traditional" family, the husband may choose only to intervene in specific areas when the differences between expectations of the wife's performance and actual outcomes surpasses a critical threshold level. In more "egalitarian" families the various roles performed, as reflected in the division of labor concerning different product areas, does not reflect delegation but rather a consensus that this is the fairest way to divide up the tasks. This consensus is encompassed in a normative framework that is determined by external forces (current cultural role shifts for women; legal developments; mass media content) as well as internal forces (joint need for competence in a marriage partner; both partners working and contributing economic resources). This framework permeates all levels of decision making, not Just the product level.

FIGURE 1

DECISIONS AT WHAT LEVEL?

Notice also that there can be upward flows in Figure 1. Since we are dealing with role performances, there is feedback from decisions made at different levels. For example, a husband who gives his wife a necklace she has always wanted (item or brand decision), not only stimulates satisfaction at the product level, but also at the family functioning level (i.e. the necklace and her satisfaction with it, represent successful fulfillment of one part of the husband role-namely affection (solidarity). This multi-level feedback would be particularly evident in decisions having strong symbolic value to family members.

Thus, researchers must specify at which level they are measuring influence as well as at which level they are measuring satisfaction with the outcomes of decisions (i.e. role performance). The latter point suggests that two stages be added to the three Davis and Rigaux (1974) focused on. These additional stages would be implementation and evaluation of a decision. Implementation and evaluation processes serve as mechanisms by which role performances are assessed and ultimately serve as benchmarks for subsequent decisions and role performances. On a broader scale, the inclusion of these stages is consistent with the bulk of research on problem solving; where processes preceding, during, and after a decision are measured (Sagasti and Mitroff, 1973).

(iv) Who do you ask?

Heer (1963) has noted that most measures of family power and family decision-making focus on the conjugal pair (i.e., the husband and wife. When various family types are considered, however, (including nuclear and extended) it becomes clear that not all relevant members are being included. The solution to this problem is akin to the use of the buying center concept (Wind, 1971) in organizational buying behavior. Rather than a priori assuming that the conjugal pair is the appropriate unit of analysis, research might focus on the identification of the other relevant dyads (where members exhibit influence attempts) involved in the decision and evaluation process.

Identification of the roles played during decision making would help identify the conditions under which it is worthwhile to consider more than the conjugal pair. For example, if a decision is dominated by the mother-father roles, pursue the mother-son/daughter or father-son/daughter roles that might also be operating in the decision. Such would not be the case if the husband-wife roles seemed to be dominating a decision.

Another way to attack the "who to ask?" problem is to use reputational and issue analysis. Using a snowball sampling approach, you ask different individuals who the most influential people are in family decision making in general (reputational analysis). You also ask what roles these people played in specific controversies or decisions that arose in the family (issue analysis). This approach provides some convergent validity by tapping different individuals' perceptions of the process. It may also lead to the inclusion of people who are not necessarily members of the immediate family but were influential in the decision making and evaluation process (e.g. grandparents, aunts-uncles, in-laws, friends, etc.).

(v) Incongruent responses

Considerations of "who to ask" ultimately lead to reconciling incongruent responses from different family members.

Davis (1976) and others have indicated that although aggregate scores may reflect a particular influence configuration, examination of specific product decisions may reveal considerable disagreement between husband and wife. Davis and Rigaux (1976) suggest that some of these differences may be attributable to either of two types of sources: modesty- either or both spouses overestimating the other's influence or underestimating their own; vanity-either or both spouses overestimating their own influence or underestimating the other's influence. Heer (1962) found that in general, husbands claim less power for themselves than their wives claim for them.

Although these represent plausible explanations for incongruencies, there are others that are equally plausible and should be tested:

a) Role Taking Accuracy: a measure of congruence could be taken to ascertain the degree to which individuals agree on the roles they are performing. It may be that the incongruency exists because one or both members are inaccurate role takers (i.e., they are not meeting the expectations of others involved in the decision) either because true disagreement exists or due to miscommunication during the decision making process.

Operationally, family congruence reflects the degree to which family members have knowledge of and agree with the expectations other members of the family have for them. In consumer decision-making, this represents the congruence that exists prior to the exchange of information and actual purchase. A high degree of congruence indicates that family members have overlapping frames of reference concerning what are the similarities and differences in expectations. [NOTE: Although there could be numerous explanations for these differences, one interesting area of inquiry concerns the similarities or differences in the ways in which information is processed by members within the dyad. Are systematic differences in information processing styles associated with lower congruence as well as greater conflict in the dyad?]

Ferraira and Winter's (1965) research indicates that families high in congruence (i.e., show higher indicated agreement as to knowledge of other members' likes, dislikes, and expectations) spent less time in reaching a decision and arrived at more appropriate decisions in terms of better fulfillment of family members' individual choices. Thus, high congruence would be associated with high levels of consumer satisfaction in contexts of joint decision making.

The interactionist framework mentioned earlier implies constant learning and modification of knowledge/ attitudes/meanings as family members participate in social exchanges with one another. Thus, congruence can improve either as a direct result of increased interaction at any point in time (higher intensity) or at the cumulative effects of interaction over a period of time (e.g. maturation). This latter point suggests that family lify cycle may be a moderator of family congruence.

Families in later stages have had a longer time to establish overlapping frames of reference and monitor expectations, particularly for decisions that are predominantly routine, a habit of nature. Congruence and subsequently satisfaction should be higher for these families as compared to families in earlier stages of the life cycle who are purchasing similar products.

However, an interesting paradox might occur in non routine, particularly "modified-re-buy" exchanges. The family's past history and interaction has established normative expectations that serve as benchmarks for current behavior. This results in a "rigidity" or stereotypical effect, where both people assume that the old rules (cognitive and social) apply in the new setting, Partial support for this hypothesis comes from Hoffman et al. (1963). He performed an experiment which revealed that group experiences with a simpler problem many times results in an inhibitory rather than facilitating effect on members' abilities to solve a similar but more complex problem. To the degree that members see particular new consumer decision problems as modifications of old problems/behaviors, then differences in perceived influence and satisfaction occur at two levels:

(i) In an objective or relative sense (i.e., compared to others making the decision), the quality of their decision in the new setting is poorer than people with less experience; they are less flexible and fewer alternatives are seen.

(ii) In a subjective sense, they assume that their partners' likes and dislikes (expectations) carry over to this new decision, when in fact the new situation elicits different expectations. These are expectations that are not predictable at the outset but occur in post-purchase evaluations since there was greater variance in the outcomes than was originally assumed using the old rules.

b) Intention vs. Behavior: the latter point on role taking accuracy, leads to a more general problem: although most of the research on family members involvement and roles attempts to measure actual behavior what is probably being measured are intentions and expectations. In a self-report measure, one spouse may be reacting to what should be while the other is reacting to what actually happened. Since there is likely to be a natural variation in this incongruency, current research may be in fact measuring two different constructs. Araji (1977) researched this potential in-congruency between prescribed attitudes and behavior. Araji splits the two (prescribed vs. actual) and builds them into the research design.

A self-administered questionnaire included items on family roles and standard demographics. Questionnaires were sent separately to individual family members. The roles measured were provider, housekeeper, child care, girls' and boys' socialization, recreation and kinship (i.e. communication with relatives). [These are based on a number of studies, most notably Nye (1974). However, Araji does not include the therapeutic and sexual roles which Nye uses in addition to these basic seven.] The results indicate that there were rather substantial numbers of married men and women experiencing role attitude-behavior incongruence. In particular, both men and women express egalitarian attitudes toward these roles, but with more women than men expressing egalitarian attitudes toward all roles. However, while men appear to express egalitarianism toward the majority of roles, this is not reflected in their role behavior. Women tend to be the major enactors of the housekeeping, boys' and girls' socialization, recreation and kinship roles; while there is a sharing of the provider and child care roles.

Subsequent research on family decision making could benefit by pursuing attitude behavior discrepancies, as well as tying the product decisions currently being used to the predominant roles being played in the family. Those used by Araji (1977) and Nye (1974) are not exhaustive but represent a reasonable first step.

It should be stressed, however, that the operationalization of these constructs needs increased attention, so that respondents' answers are not merely artifacts of a researcher's demand bias.

Self-Report Vs. Observational Methods

The latter point indicates the need to develop a greater variety of methods to study family role structures and decision making.

Traditional approaches to studying these areas makes the distinction between self-report and observational methods. Both methods have strengths and weaknesses.

Kenkel (1961) has stated that self-reports assume the following:

- that people know the relative influence they have

- are willing to admit it

- are able to recall with accuracy how influence was distributed.

Safilios-Rothschild (1970) has also noted that final outcome self-reports capture only one part of the phenomenon. She found that in Husband-Dominant decision making, three quarters of the women used some type of verbal technique in influencing their husbands but officially let them "make the decisions". Non verbal techniques (facial expression, etc.) are other communications that would occur in addition to verbal cues. She also found that 70% of husbands in Wife-Dominated decisions used influence techniques, and even after the wife "made the decision," 23% did what they wanted anyway (i.e. the product purchase reflected the husband's preference not the wife's).

Observational techniques are viewed as a way of getting around some of these problems as well as providing greater "realism' to the situation. This becomes particularly important when research questions focus on how patterns of social exchanges reflect consensual goals in some situations while others reflect conflicting goals. The strategies of problem solving, persuasion and bargaining that occur within these patterns are particularly suited to observational techniques.

For example, Bales IPA (1950) would measure husband-wife exchanges in twelve dimensions: solidarity, tension, release, agrees, gives suggestions, gives opinions, given orientation, asks for orientation, asks for opinion, asks for suggestions, disagrees, shows tension and shows antagonism.

Olson (1969) used Strodtbeck's revealed differences method when couples disagreed on particular questions. Questions are given to each member separately, responses are compared, and differences are pinpointed. Family members are presented with these differences and asked to explain and resolve the differences. These exchanges are measured by recording the interaction and ascertaining the influence processes involved as well as the final out coma.

Webb (1978) is developing "unsolicited protocols" to study the communication of spouses by placing tape recording equipment in homes of volunteers. For a two week period voice sensitive equipment is activated automatically when conversation occurs. Responses are coded in eight categories: problem information or feelings about problems, mindreading (attributing feelings to the other spouse), proposing a solution, communication talk, agreement, disagreement, summarizing other, summarizing self.

However, the stated "realism" of observational techniques has been criticized for a number of reasons:

- you are viewing public behavior and social desirability biases may enter.

- many of the situations presented to families have been criticized as being too atypical or unimportant, thus seriously limiting external validity.

- Olson (1969) found that in the lab context there tends to be more (vs. self reports) disagreement between spouses, less efficiency at decision making, and less emotionality registered.

- Kenkel (1961) has shown that the sex of the observer greatly influences the participants behavior. It was indicated that when the observer was a woman, wives tended to take more active and powerful roles.

- Safilios-Rothschild (1969) suggests observational techniques are limited:

a. when the behavior was of a highly intimate nature

b. because of considerations of optimal timing (e.g. when she's in a good mood or he got a raise)

c. because the application of a technique requires special tasks (such as cooking a special meal, taking her out to dinner)

d. because a long time and repeated application were required (e.g. she was dissatisfied but had to keep nagging him to do something about it).

Given the limitations and benefits of each method as well as the exploratory nature of many family role structure studies, it is suggested that further research requires a mixture of both approaches. This need is underscored by the nature of the four necessary conditions for identifying roles and role structures:

- roles refer to behavior performances,

- roles imply interaction between two or more individuals,

- roles are rule-based (derived from normative expectations - cultural group, and individual),

-roles indicate a bundle of attributes or behaviors.

SUMMARY

The family as a unit of analysis represents a common sense but complex way to view consumer decision making. This perspective involves viewing the family both as a statistical group as well as a primary social group. The latter view underscores a concern not only for identifying a diversity in family role structures at an aggregate level but also focusing on the processes of communication, interaction and influence that occur within specific family role structures. To do so requires increased attention to the definition of roles; specification of decisions as they relate to roles and at what level; using roles to determine who to talk to; viewing response incongruency as a function of role taking inaccuracy and an attitude behavior discrepancy; and integrating both self-report and observational methods.

REFERENCES

Araji, S. K. (1977), "Husbands' and Wives' Attitude -Behavior Congruence on Family Roles," Journal of Marriage and the Family, Vol. 39, 309-320.

Babble, E. R., (1973), Survey Research Methods, Belmont, CA: Wadsworth.

Bagozzi, R. P. and VanLoo, M. F., (1978), "Fertility as Consumption: Theories from the Behavioral Sciences," Journal of Consumer Research, March, 199-228.

Bahr, S. J., (1973), "The Internal Consistency of Blood and Wolfe's Measure of Conjugal Power: A Research Note," Journal of Marriage and the Family, 35, 293-295.

Bales, R. F., (1950), Interaction Process Analysis, Reading, Mass, :Addison-Wesley.

Blood, R. and Wolfe, D., (1960), Husbands and Wives: The Dynamics of Married Living, Glencoe: The Free Press.

Blood, R., Hill, R., Michel, A., and Safilios-Rothschild C., (1970), "Comparative Analysis of Family Power Structure: Problems of Measurement and Interpretation," in Hill, R., and Konig, R. (eds.) Families in East and West Paris: Maitan.

Broderick, C. B., (1971), "Beyond the Five Conceptual Frameworks: A Decade of Development in Family Theory," Journal of Marriage and the Family, 33, 139-59.

Cunningham, I., and Green, R., (1974), "Purchasing Roles in the U.S. Family," Journal of Marketing, 38, 61-81.

Davis, H. L. and Rigaux, B., (1974), "Perceptions of Marital Role in Decision Processes," Journal of Consumer Research, July, 51-61.

Eshelman, J. R., (1975), The Family: An Introduction, Boston: Allyn and Bacon.

Ferber, R., and Lee, L.C., (1977), "Husband-Wife Influence in Family Purchasing Behavior," Journal of Consumer Research, June, 19-28.

Ferraira, A. J., and Winter, W. D., (1965), "Family Interaction and Decision Making," Archives of General Psychiatry, 13, 214-223.

Granbois, D. H., and Summers, J. O., (1975), "Primary and Secondary Validity of Consumer Purchase Probabilities,'' Journal of Consumer Research, March, 31-8.

Granbois, D. H., Summers, J. O., and Frazier, G. L., (1977), "Correlates of Consumer Expectation and Complaining Behavior,'' in R. L. Day (ed.) Consumer Satisfaction, Dissatisfaction and Complaining Behavior, Bloomington, Ind.: Indiana University, 18-25.

Green, R., and Cunningham I., (1975), "Feminine Role Perception and Family Purchasing Decisions," Journal of Marketing Research, August, 325-32.

Heer, D. M., (1962), "Husband and Wife Perceptions of Family Power Structure," Marriage and Family Living, February, 65-67.

Heer, D. M., (1963), "The Measurement and Bases of Family Power: An Overview," Marriage and Family Living, 25, 133-39.

Heffring, M. P., (1978), "Consumer Satisfaction/Dissatisfaction: The Family Perspective," Chicago, Consumer Satisfaction/Dissatisfaction Conference.

Hempel, D. J. and Ayal, I., (1977), "Transition Rates and Consumption System: A Conceptual Framework for Buyer Behavior in Housing Markets," in Consumer and Industrial Buyer Behavior, Arch G. Woodside, J. Sheth, P. D. Bennett (eds.), New York: North Holland Publishers, 201-218.

Herbst, P., (1952), "The Measurement of Family Relationships,'' Human Relationships," Human Relations, 5, 3-35.

Hoffman, L. , Burke, J. R., and Maier, N. R. , (1963), "Does Training With Differential Reinforcement on Similar Problems Help in Solving a New Problem," Psychological Reports, 13, 147-154.

Hunt, H. K., (1977), "CS/D: Bits and Pieces," in R. L. Day (ed.) Consumer Satisfaction, Dissatisfaction and Complaining Behavior, Bloomington, Ind.: Indiana University, 38-42.

Hill, R., and Hansen, D. A., (1960), "The Identification of Conceptual Frameworks Utilized in Family Study," Marriage and Family Living, 22, 299-311.

Jacoby, J., (1978), "Consumer Research: A State of the Art Review," Journal of Marketing, 42,87-96.

Kenkel, W.F., (1961), "Family Interaction in Decision Making On Spending," in Foote, N. (ed.) Household Decision Making, New York: New York University Press,

Kuhn, T. S., (1962), The Structure of Scientific Revolutions, Chicago: University of Chicago Press.

Mason, J., and Himes, S., (1973), "An Exploratory Behavioral and Socio-Economic Profile of Consumer Action About Dissatisfaction With Selected Household Appliances,'' Journal of Consumer Affairs, Winter, 121-27.

Miller, J., (1977), "Data Reduction Techniques and the Exploration of Satisfaction Segments," in R. L. Day (ed.) Consumer Satisfaction, Dissatisfaction and Complaining Behavior, Bloomington, Ind.: Indiana University, 102-114.

Nadel, S. F. (1974), The Theory of Social Structure, London: Cohen and West.

Nye, F. I., (1974), "Emerging and Declining Family Roles," Journal of Marriage and the Family, 36, 238-45.

Olson, D. H., (1969), "The Measurement of Family Power by Self Report and Behavioral Methods," Journal of Marriage and the Family, 31, 545-550.

Rigaux-Bricmont, B., (1978), "Explaining Marital Influences in The Family Economic Decision Making," in AMA 1978 Educators Proceedings, S. C. Jain (ed.), American Marketing Association, 126-129.

Robinson, P. J., and Faris, C. W., (1967), Industrial Buying and Creative Marketing, Boston: Allyn and Bacon.

Safilios-Rothschild, C., (1969), "Patterns of Familial Power and Influence," Sociological Focus, 2, 7-19.

Safilios-Rothschild, C. , (1970), "A Study of Family Power Structure: A Review 1960-1969," Journal of Marriage and the Family, 32, 539-552.

Sagasti, F. R., and Mitroff, I. I., (1973), "Operations Research From the Viewpoint of General Systems Theory," Omega, 695-709.

Straus, M. A., and Tallman, I., (1971), "SINFAM: A Technique for Observational and Experimental Study of Families," in J. Aldores (ed.) Family Problem Solving, Hinsdale: The Dryden Press, 381-438.

Wall, M., Dickey, L., and Talarzyk, W., (1977), "Predicting and Profiling Consumer Satisfaction and Propensity to Complain," in R. L. Day (ed.) Consumer Satisfaction, Dissatisfaction and Complaining Behavior, Bloomington, Ind.: Indiana University Press, 91-101.

Wallendorf, M., (1978), "Social Roles in Marketing Contexts ," American Behavioral Scientist, 21, 571-582.

Webb, P. H., (1978), "A New Method for Studying Family Decision Making," Journal of Marketing, 42, p.12.

Weik, K., (1971), "Group Processes, Family Processes, and Problem Solving," in J. Aldous (ed.) Family Problem Solving, Hinsdale: The Dryden Press, 3-32.

Wind, Y., (1971), "A Reward Balance Model of Buying Behavior in Organizations," in G. Fish (ed.), New Essays in Marketing Theory? Boston: Allyn and Bacon.

Wolgast, E. , (1958), "Do Husbands or Wives Make the Purchasing Decisions," Journal of Marketing, 23, 151-58.

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