Data regarding consensus and mental health were collected from 252 women, their husbands, and an adolescent child who were members of three-generation households. Analyses indicate that consensus is best represented by six separate constructs and mental health by three separate constructs. Mental health of individual family members was differentially predicted by indicators of consensus.
Understanding people within families has been a goal of both psychologists and sociologists. Neither discipline, however, has done a particularly good job of matching theory with methodological approaches, resulting in a knowledge base that is generally characterized by more questions than answers. The goal of this article is to discuss, from both a conceptual as well as a methodological perspective, the constructs of family consensus and mental health, and to examine the way in which these two constructs relate to one another.
Larsen and Olsen (1990) define "family study" as one that "systematically integrates the family perspective from the conception of the project through to the final analysis" (p. 20). One issue that has come to the attention of researchers is the extent to which we can learn about families from the report of only one of its members. Thompson and Walker (1982) suggest that there are situations when data from one individual in the family can adequately represent the family. Data collected from individuals qualify as research about families if the intention is to use an individual's report as either: (a) an objective reality; implying that the report is independent of the individual's view, such as report of the number of family members, length of time a marriage has existed, or the presence/absence of characteristics that can be corroborated by an outside observer, or (b) a subjective individual reality that is interpreted as one family member's perception of himself, her family, or other family members. If however, the intention is to use one family member's subjective perceptions to characterize the feelings of a whole family, or to try to represent a family's reality beyond the confines of its physical and demographic characteristics, then gathering information from only one family member is insufficient (Pruchno, 1989).
Safilios-Rothschild's (1969) eloquent criticisms of family sociology as "wives' family sociology," as well as similar comments from other scholars (Booth & Welch, 1978; Cromwell & Olson, 1975; Larson, 1974; Olson, 1977; Thomas & Colonico, 1972), have been answered by an increase in the number of studies that collect data from more than one family member. However, as Larsen and Olson (1990) have warned, "just as it is dangerous to assume that one member can adequately represent the family's reality, it is equally problematic to conclude that the acquisition of several members' data can provide a more valid reality" (p. 22). In fact, responding to Safilios-Rothschild's (1970) lament, Larsen and Olson (1990) point out that there is neither theoretical justification nor empirical findings that indicate that husbands, sons, daughters, or a family composite can provide any more insights than those provided by wives alone. Certainly, it can be argued that collecting data from multiple family members results in a richer, more intricate description of what is going on within the family, yet the meaning and usefulness of these reports must be established at the theoretical rather than methodological level.
Central to the goal of increasing our understanding of families is making and respecting the distinction between characteristics of individuals and characteristics of families. Depending on the theoretical assumptions one makes about this issue are decisions regarding research design, data analysis, and interpretation. One of the salient problems that has plagued students of the family is a lack of respect for unit parity. As such, characteristics of individuals and characteristics of families are frequently confounded (Pruchno, 1989; Thompson & Walker, 1982). …