Advancing a family perspective in health research: models and methods.

In this essay we review and respond to four problems that have impeded the application of family theory to health research: the difficulty in translating family concepts to empirical health research; the confusion and uncertainty in measuring "wholes" and "parts" of a family in a single health project; the apparent conflict between circular-causal and linear perspectives in model building; and the putative, implicit "violation" of a family perspective by the employment of certain data management and design strategies. We argue that current, complex conceptualizations of the family have to be matched with equally complex research designs and methods, and that techniques are available to reach these goals. We distinguish between theoretical models and research methods, and review several strategies, design issues, and alternatives for data analysis. Throughout, we point to 1) the need to maintain consistency among theory, construct, and indicator, 2) the employment of multidimensional family assessment in health research, and 3) the use of readily available but infrequently used methods and procedures for appreciating the richness of multivariate family data.