An Exploratory Study of Family Decision Making Using a New Taxonomy of Family Role Structure

Irene Raj Foster, Indiana University
Richard W. Olshavsky, Indiana University
ABSTRACT - This paper presents the results of an exploratory study on family decision making that uses a new extension of the existing taxonomy of family role structure. The primary objective of the study was to determine the feasibility of measuring the occurrence of the various types of group decision making structures described in the new taxonomy. A secondary objective was to compare the distribution of the observed structures in a restaurant selection decision on special and nonspecial dining occasions. The results demonstrate: 1) the feasibility of measuring group structures using a survey approach, 2) that structures differ significantly across families, and 3) that no significant difference in structures occurs across situations.
[ to cite ]:
Irene Raj Foster and Richard W. Olshavsky (1989) ,"An Exploratory Study of Family Decision Making Using a New Taxonomy of Family Role Structure", in NA - Advances in Consumer Research Volume 16, eds. Thomas K. Srull, Provo, UT : Association for Consumer Research, Pages: 665-670.

Advances in Consumer Research Volume 16, 1989      Pages 665-670

AN EXPLORATORY STUDY OF FAMILY DECISION MAKING USING A NEW TAXONOMY OF FAMILY ROLE STRUCTURE

Irene Raj Foster, Indiana University

Richard W. Olshavsky, Indiana University

ABSTRACT -

This paper presents the results of an exploratory study on family decision making that uses a new extension of the existing taxonomy of family role structure. The primary objective of the study was to determine the feasibility of measuring the occurrence of the various types of group decision making structures described in the new taxonomy. A secondary objective was to compare the distribution of the observed structures in a restaurant selection decision on special and nonspecial dining occasions. The results demonstrate: 1) the feasibility of measuring group structures using a survey approach, 2) that structures differ significantly across families, and 3) that no significant difference in structures occurs across situations.

It is generally recognized that the "family" rather than the "individual" should be the basic unit of analysis in the study of consumer behavior (Grashof and Dixon 1980). Studying family decision making is difficult, however, in as much as it entails the synthesis of the goals, preferences, strategies and perceptions of all family members who are part of the decision making group. Furthermore, important differences characterize families and important changes occur within families in terms of the family's composition, structure, and tasks performed by various family members, both over time and over the stages of the decision process (Davis and Rigaux 1974; Davis 1976). In spite of these difficulties, researchers in the area of family decision making recognize the importance of "exploring different methodologies for measuring family role structure, factors predictive of family role structure, and the effects of family role structure on household buying behavior" (Szybillo and Sosanie 1977).

The purpose of this paper is to describe the results of an exploratory study designed to: 1) ascertain the types and relative frequency of occurrence of the role structures identified in a new, more detailed taxonomy of family role structure and 2) test one hypothesis concerning the impact of a situational variable on the types of roles structures that occur. We begin with a brief review of past research in the area of family decision making in order to point out the major contributions and limitations of the existing taxonomy. This will be followed by a description of the new taxonomy and a review of its major advantages for the study of family decision making. Finally, we will describe some directions for further research on family decision making and discuss some limitations of the present research.

THE (PRESENT) TAXONOMY OF FAMILY ROLE STRUCTURE

The taxonomy of role structures that has been widely adopted is the one used by Davis (1976). This taxonomy was originally developed by Herbst (1954) and has subsequently been used by many other sociologists and marketers. (We did look at other group decision making studies and models in organizational behavior, economics and psychology; however, these will not be examined or discussed in greater detail here as this paper is not taking a normative but a descriptive approach to family decision making.) According to Herbst's taxonomy, there are four types of family role structures: 1) husband dominant, 2) wife dominant, 3) autonomic, and 4) syncratic. Husband dominant and wife dominant families are defined as those where the husband or wife plays the dominant role in making the purchase decision. On the other hand, within any sample of families, if more than 50% make a decision jointly, the decision is classified as "syncratic;" if less than 50% make a decision jointly, the decision is classified as "autonomic" (p. 53, Davis and Rigaux 1974).

Studies that have employed Herbst's taxonomy have largely been concerned with identifying the conditions under which the various role structures occur (see, for example, Davis 1970; Davis and Rigaux 1974; Green and Cunningham 1975; Hempel 1974; Hempel 1975; Munsinger, Weber and Hansen 1975; Sheth and Cosmas 1975; Wolgast 1958; Woodside 1972; Woodside 1975).

Past research has shown that whether a decision is "syncratic" or "autonomic" depends upon two types of variables. The first type consists of variables such as family life cycle, socioeconomic status and life styles while the second type consists of variables such as perceived risk, importance of purchase, time pressure and situational contingencies (Sheth and Cosmas 1975). In the study conducted by Sheth and Cosmas (1975), families mentioned that the greater competence of one partner, preference for dividing responsibilities in household management, greater importance of the decision to one of the spouses, being too busy to decide together and peer group norms were primarily responsible for "autonomic" decisions. They also found that decisions tended to be "autonomic" for households with either no or grownup children, in low or middle socio-economic classes and with the wife in some blue-collar occupation. Granbois (1971) points out that social class, education, income, family size and the wife's employment status are all factors that influence whether a family makes "syncratic" or "autonomic' decisions.

When joint decision making does take place, previous research (Blood and Wolfe 1960; Davis 1970; Davis and Rigaux 1974; Green and Cunningham 1975; Kollat, Engel and Blackwell 1970; Sharp and Mott 1956; Wolgast 1958; Woodside 1975) has found that there is a difference in family role structures and member decision roles across types of products and services. In addition, within different product categories, role structures vary by the nature of the decision to be undertaken or the phase of the decision process (Burns 1977; Davis 1970; Davis and Rigaux 1974; Hempel 1974; Munsinger, Weber and Hansen 1975; Starch 1958; Szybillo and Sosanie 1977).

AN EXTENSION TO THE TAXONOMY OF FAMILY ROLE STRUCTURE

While Herbst's taxonomy has been widely used by researchers it's main drawback is that only two of the four types of role structures - autonomic and syncratic - capture group decision making (i.e., decision making in which two or more family members interact to make the decision). More importantly, however, a decision is classified as involving autonomic or syncratic role structures dependent upon the percentage of the sample of families making a- joint decision. Thus, while the taxonomy facilitates the examination of when and under what conditions family decision making takes place, it does not facilitate a deeper understanding of the processes of family (joint) decision making or reveal the impact of group decision making structure on the outcome and process of family decision making.

This lacuna in the taxonomy was originally pointed out by Olshavsky and King (1984; 1985) and in an attempt to help overcome this important deficiency, Olshavsky and King (1985) proposed an interesting and significant extension to the existing taxonomy of family role structure. Their extension of the taxonomy is based upon the "distributed processing" concept that is central to network design (Martin 1981). It is also directly related to the much earlier work done by organizational theorists on group decision making (e.g., Bavelas 1950; Leavitt 1958).

According to the new taxonomy, to make a choice, a family can adopt any one of five basic "structures:" parallel, hierarchical, ring, and star (with two separate forms, A and B). (See Figure). An almost endless array of hybrid configurations may be formed from two or more of the basic structures. Further, within each structure adopted, a family can "assign" a particular task or a component of a task to one or more of the family members that comprise the decision making group on that occasion.

Parallel

In this arrangement, two or more members of the family unit work on the same decision simultaneously and independently. Assignment is not an issue in this arrangement. No communication occurs among family members.

Hierarchical

In this arrangement, two or more members of the family unit are ranked in terms of their decision making capabilities. The decision is assigned first to the member of lowest rank. If that member is unable to make the assigned decision, the partially made decision is passed on to the member ranked next highest. The passing of the progressively more processed decision continues until the decision is either made or the number of members in the family is exhausted. Thus, the processing of the decision as well as the allocation of the decision to members is sequential in this arrangement. Communication occurs only between adjacent members in the hierarchy.

Ring

In this arrangement, no ranking of family members occurs; the members that comprise the ring are judged to have similar decision making capabilities. However, each member may be thought of as a "specialist". Accordingly, decisions or different aspects of the same decision are assigned to the most appropriate member in the ring (i.e., on the basis of fit between decision to be made and member specialization). In this arrangement, the decisions or component decisions are made simultaneously and/or sequentially by different members. Communication can occur among any members of the ring.

Star

In this arrangement, one member is assigned the task of coordinating the efforts of all other members in the family unit. Decisions or components of decisions are assigned to the coordinator and/or to one or more of the other members on the basis of their decision making capabilities. In Form A, communication among family members can only occur indirectly through the coordinator. In Form B, there can be direct as well as indirect communication among members.

THE EMPIRICAL STUDY

Research Objectives

The new taxonomy makes clear that there are many different ways in which families could structure themselves in the case of actual group decision making. To our knowledge, none of the many studies of family decision making has attempted to study role structure at this-level of detail. Therefore, the primary purpose of this exploratory research was to determine the feasibility of measuring the specific structures adopted and, if feasible, to determine the types and relative frequency of occurrence of the role structures identified in the new, more detailed taxonomy of family role structure. Further, it is clear that the specific structure adopted may depend on the conditions under which the decision is being made. Therefore, a secondary purpose of this exploratory study was to test a hypothesis concerning the impact of a situational variable on the types of group decision making structures that occur (Belk 1984).

There has traditionally been a dichotomy in the situational influence literature with some researchers supporting psychological measurement of situations (Lutz and Kakkar 1975) and others supporting objective measurement of the same (Belk 1975a). For the purposes of this study the latter approach was chosen as "it removes the idiosyncracies of perception which may otherwise limit aggregation and manipulation of consumer situations" (Belk 1975b). Thus, situations were defined objectively in terms of task definition features in this study. However, it must be noted that if a study is to be of a conclusive nature or if results are to be generalized, then the research design must incorporate both types of measurement.

FIGURE

ALTERNATIVE STRUCTURES WITHIN A MULTI-INDIVIDUAL CONSUMER UNIT

Restaurant choice was selected as the task of interest in this study. Past research (e.g., Miller and Ginter 1979; Szybillo and Sosanie 1977) has shown that family decision making is likely to occur for restaurant choice and that type of restaurant selected does vary across situations. Within this choice task, "special" versus "nonspecial" dining occasions were chosen as the situational variable of interest. A special occasion was defined as an occasion such as a birthday or an anniversary while a nonspecial occasion was defined as an occasion when the cook of the family was just too busy or tired to cook or the family wanted some variety in their meals.

METHOD

Respondents

A convenience sample of 75 families (a family being defined as a household unit with both parents and one or more children living at home) participated in this study. The sample consisted of families living in student housing on campus and families from several local churches who chose to participate in the study.

Questionnaire Design

The questionnaire consisted of three parts. Parts I and II described a recent special and a recent nonspecial dining occasion and elicited information from respondents as to how the family had made a restaurant selection decision on that occasion, which group decision making structure had been adopted in making that decision (if a group decision had occurred), and reasons for selecting that structure. Each of the basic structures was described in detail using a combination of text, illustrations and several examples unrelated to restaurant choice. Several versions of the questionnaire were pretested on small convenience samples to ensure the clarity of the description of each of the basic structures and to ensure that respondents were clear that the two dining situations were different and were treating them as such. After several rounds of pretesting, respondents appeared to understand and to readily select various structures for each of the various illustrative choice situations used. Part III elicited information on age, educational level, occupation and income.

The questionnaire was developed using two different formats. Part I discussed the special dining occasion in half the questionnaires and the nonspecial dining occasion in the other half. This was done to control for order effects.

Procedure

Both husband and wife were asked to fill out the questionnaire jointly. Each family that sent in a completed questionnaire received $5.00.

Hypotheses

Since some group decision making structures are more appropriate than others for families to adopt when making a restaurant selection decision, it was expected that, in general, the structures would not be equally likely to occur. Due to the large differences in degree of interaction and participation possible in each structure, it was expected that for this product decision (i.e., restaurant selection), the structures that would most often be adopted by families would be the ring (i.e., the most "group" oriented structure), star B, star A and hierarchical structures in descending order of usage. It was expected that the parallel structure would seldom, if ever, be adopted since there is, by definition, no interaction among members even though each participates (independently) if such a structure is adopted. In addition, it was expected that the five structures would not be equally likely to occur either in the special or nonspecial situations taken alone. Thus, our first hypotheses, expressed in null form are:

H1a: The five structures are all equally likely to occur for the restaurant selection decisions (special and nonspecial).

H1b: The five structures are all equally likely to occur for a restaurant selection decision on a special occasion.

H1c: The five structures are all equally likely to occur for a restaurant selection decision on a nonspecial occasion.

It was also expected that the situational variable - type of occasion - would have an impact on the group decision making structures adopted. Specifically, it was expected that the hierarchical and star A rather than star B or ring would be adopted more frequently on the special occasion as the family would be concerned about making the dining occasion a success and these two structures could ensure that the relevant group members' desires would be taken into account. Thus, our second hypothesis, in null form, is:

H2: There will be no difference in the structures chosen on the special versus the nonspecial occasion.

Results

The response rate was 76%; 56 questionnaires were fully completed and mailed back.

The Table presents the observed frequency of occurrence of each of the structures for special, nonspecial and combined. The "other" category refers to those instances in which a single person in the family made the restaurant selection decision for the entire family.

Given the nonmetric nature of the data, the hypotheses were tested using Kolmogorov-Smirnov one and two-sample tests (Siegel 1956). The Kolmogorov-Smirnov one-sample test showed that H1a was rejected at a significance level of .01. If the family adopted a group decision making orientation, the most commonly reported structure was the ring, the next most common was Star B, then Star A; Hierarchical and Parallel were tied for last position. These results clearly suggest that there are large differences across families in the structures adopted for restaurant choice. Similarly, using Kolmogorov-Smirnov one-sample tests, it was found that H1b and H1c were both rejected at a significance level of .01.

H2 was not rejected using a Kolmogorov-Smirnov two-sample test. A chi-square test showed that H2 could be rejected at a significance level of .01. However, this test is less powerful than the Kolmogorov-Smirnov two-sample test. Thus, the distribution of structures across the two occasions is not significantly different. H2 may not have been rejected due to our small sample size. A chi-square analysis of the types of reasons given for selecting the ring structure showed a significant relationship between the reason "because this organization assures that everyone gets what he/she wants" and choice of the ring structure at the 0.001 significance level. A chi-square analysis examining reasons for the choice of structures on the special dining occasion showed that the reason "in order to train young children in the family to make such decisions" was significantly related to the choice of the ring and star B structures at the 0.003 significance level while the reason "because this organization assures that everyone gets what he/she wants" was significantly related to the choice of the ring and star A structures at the 0.05 significance level. Examination of the nonspecial dining occasion shows that the reasons "because this organization assures that everyone gets what he/she wants" and "because we think the best restaurant is selected this way" were significantly related to the choice of the ring and star A structures at the 0.05 significance level while the reason "because we have always done it this way" was significantly related to the choice of the ring and star B structures at the 0.05 significance level.

The "other" category (i.e., those instances in which a single person in the family made the restaurant selection decision for the entire family) was observed more often on the special dining occasion (16 times) than on the nonspecial dining occasion (5 times). In the case of the special dining occasion, one person made the restaurant selection decision because:

1) "the special occasion was in honor of that person" (9 out of 16 times),

2) the person was either paying for the treat or it was his/her idea to go to a particular restaurant and everyone in the family liked the choice of restaurant (5 out of 16 times), and,

TABLE

FREQUENCY OF REPORTED USAGE OF EACH TYPE OF GROUP STRUCTURE FOR SPECIAL AND NONSPECIAL OCCASIONS

3) that person was most knowledgeable about restaurants (2 out of 16 times).

In the case of the nonspecial dining occasion, one person made the restaurant selection decision for the entire family because that person usually made restaurant selections for the rest of the family. In fact, the survey revealed that this individual was in charge of or took charge of other types of decisions for the family as well.

A CROSSTABS chi-square analysis was also run on a series of demographic variables such as size of the family, income, education and children's ages. However, none of the demographic variables were significantly related to the particular structures chosen for family decision making.

DISCUSSION

Based upon this exploratory study, it appears that measurement problems are surmountable and that families do adopt structures such as those described in Olshavsky and King's proposed extension of Herbst's taxonomy of family role structure. Our preliminary results show that specific group decision making structures adopted vary considerably across families with the ring structure being the most popular. These results suggest that our ability to understand family decision making will remain limited until we empirically determine the types of structures actually adopted by families in different choice domains. It would certainly be interesting to examine how results will differ when husbands and wives are each requested to complete a questionnaire on their own instead of completing one jointly.

Given these encouraging preliminary results, we may now examine more systematically issues such as why families adopt these structures, how tasks are assigned once structures are adopted, and what the implications are for brand choice for each of the different role structures. Further, given knowledge of the structure adopted, the stage is set for an even more detailed study of family decision making at the level of the specific information processing rule (e.g., conjunctive or disjunctive or lexicographic) adopted by each member of the family participating in the decision.

Limitations of the Study

While this exploratory study generates many interesting results, some of the results have to be accepted with caution as sample size is small (n = 56). Also, for robust chi-square results the expected cell frequencies should be >= 5. Due to the small sample size and the fact that there were many categories for variables such as income and education, a number of the results, especially those examining demographic variables, turned out to be nonsignificant.

A second limitation of the study is the use of the survey method to examine the various structures that families adopt when making decisions. While this method is extremely useful for an exploratory study, self reports and retrospective reports of the structures adopted when making the restaurant selection decision may have produced results biased in favor of the ring structure. The ring structure may have been-reported as the structure the family adopted most often because- the ring structure, by its very nature, suggests that the family that uses this structure is a happy, harmonious family where no one person dominates the decision making process and everyone has a say in the family's choice of a product. Measurement errors of this type could be reduced by use of an observational method wherein the researcher classifies the types of structures actually used by subjects in a laboratory study involving a simulated choice task.

REFERENCES

Bavelas, A. (1950),"Communication Patterns in Task-Oriented Groups," Journal of Acoustical Society of America, 22, 725-730.

Belk, R.W. (1974),"An Exploratory Assessment of Situational Effects in Buyer Behavior", Journal of Marketing Research, 11 (May), 156-163.

Belk, R.W. (1975a)," The Objective Situation as a Determinant of Consumer Behavior", in Mary Jane Schlinger (ed.), Advances in Consumer Research, Vol. 2, Chicago: Association for Consumer Research.

Belk, R.W. (1975b),"Situational Variables and Consumer Behavior", Journal of Consumer Research, 2, 157-164.

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

Burns, Alvin C. (1977),"Husband and Wife Purchase Decision-Making Roles: Agreed, Presumed, Conceded and Disputed," in Advances in Consumer Research, Vol. 4, ed. William D. Perreault, Jr., Ann Arbor, MI: Association for Consumer Research, 50-55.

Davis, Harry L. (1970),"Dimensions of Marital Roles in Consumer Decision Making," Journal of Marketing Research, 7 (May), 168-177.

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

Davis, Harry L. and Benny P. Rigaux (1974),"Perception of Marital Roles in Decision Processes," Journal of Consumer Research, 1(June), 51-62.

Granbois, Donald H. (1971),"A Multi-Level Approach to Family Role Structure Research," in Proceedings of the Second Annual Conference, (ed.) David M. < Gardner, Ann Arbor, MI: Association for Consumer Research, 99-107.

Grashof, John R. and Donald F. Dixon (1980),'The Household: The "Proper" Model for Research into Purchasing and Consumption Behavior," in Advances in Consumer Research, Vol. 7, (ed.) J.C. Olson, Ann Arbor, MI: Association for Consumer Research, 489-491.

Green, Robert T. and Isabella C. M. Cunningham (1975),"Feminine Role Perception and Family Purchasing Decisions," Journal of Marketing Research, 12 (August), 325-332.

Hempel, Donald J. (1974),"Family Buying Decisions: A Cross-Cultural Perspective," Journal of Marketing Research, 11 (August), 295-302.

Hempel, Donald J. (1975),"Family Role Structure and Housing Decisions," in Advances in Consumer Research, Vol. 2, (ed.) Mary Jane Schlinger, Ann Arbor, MI: Association for Consumer Research, 71 -80.

Herbst, P. G. (1954), "Conceptual Framework for Studying the Family," in O. A. Oeser and S. B. Hammond, (eds.) Social Structure and Personality in a City, New York: Macmillan.

Kollat, David T., James F. Engel and Roger D. Blackwell (1970),"Current Problems in Consumer Behavior Research," Journal of Marketing Research, 7 (August), 327-332.

Leavitt, Harold J. (1958), Managerial Psychology. The University of Chicago Press: Chicago.

Lutz, R. J. and P. K. Kakkar (1975),"The Psychological Situation as a Determinant of Consumer Behavior", in Mary Jane Schlinger (ed.), Advances in Consumer Research, Vol. 2, Chicago: Association for Consumer Research.

Martin, James (1981). Design and Strategy for Distributed Data Processing. Englewood Cliffs, N. J.: Prentice-Hall.

Miller, K. and J. Ginter (1979), "An Investigation of Situational Variation in Brand Choice Behavior and Attitude," Journal of Marketing Research, February, pp. 111-25.

Munsinger, Gary M., Jean E. Weber and Richard W. Hansen (1975), "Joint Home Purchasing Decisions by Husbands and Wives," Journal of Consumer Research, 1 (March), 60-66.

Olshavsky, Richard W. and Maryon F. King (1984),"The Role of Children in Household Decision Making: Application of a New Taxonomy of Family Role Structure," in Changing Households, (eds.) Mary Lou Roberts and Lawrence H. Wortzel, Ballifger Publishing Company, 41-52.

Olshavsky, Richard W. and Maryon F. King (1985),"A New Classification of Family Role Structure, in Proceedings, Annual Convention of the American Psychological Association, Division 23, (ed.) David C.Stewart, 33-34.

Sharp, H. and P. Mott (1956),"Consumer Decisions in the Metropolitan Family," Journal of Marketing, 21 (October), 149-156.

Sheth, Jagdish N. and Stephen Cosmas (1975),'Tactics of Conflict Resolution in Family Buying Behavior," Faculty Working Paper no. 271, College of Commerce and Business Administration, University of Illinois at Urbana-Champaign.

Siegel, Sidney (1956) "Nonparametric Statistics for the Behavioral Sciences. McGraw-Hill Book Company, Inc: New York.

Starch, D. and staff (1958),"Male vs. Female: Influence on the Purchase of Selected Products as Revealed by an Exploratory Depth Interview Study with Husbands and Wives." New York: Fawcett Publications.

Szybillo, George J. and Arlene Sosanie (1977),"Family Decision-Making: Husband, Wife and Children," in Advances in Consumer Research, Vol. 4, (ed.) William D. Perreault, Jr., Ann Arbor, MI: Association for Consumer Research, 46-49.

Wolgast, Elizabeth (1958),"Do Husbands or Wives Make the Purchasing Decisions?," Journal of Marketing, 23 (October), 151-158.

Woodside, Arch G. (1972),"Dominance and Conflict in Family Purchase Decisions," in Advances in Consumer Research, Proceedings of the Third Annual Conference, (ed.) M. Venkatesan, Ann Arbor, MI: Association for Consumer Research, 650-659.

Woodside, Arch G. (1975),"Effects of Prior Decision Making, Demographics and Psychographics on Marital Roles for Purchasing Decisions," in Advances in Consumer Research, Vol.2, (ed.) Mary Jane Schlinger, Chicago: Association for Consumer Research, 81-91.

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