Book Review: Analyzing Decision Making: Metric Conjoint Analysis

The author intends the monograph to serve as " . an introduction to theory and methods for studying customer decision processes" (p. 7) for " ... researchers who want to solve practical problems with a sound theoretical approach that has had considerable empirical success. . . . " (p. 8). As the title implies, the book is about (metric) conjoint analysis. However, before discussing aspects of research design and data analysis, Louviere introduces the theory of information integration. Though the discussion of information integration is brief, it is useful for readers who might otherwise be inclined to start with an ad hoc specification of a model to represent consumer choices or preferences. The author mentions the link between physical characteristics and attribute evaluations, but provides little guidance to help researchers decide whether to use physical (objective) characteristics or attribute (subjective) descriptions or some of both categories for a study. The attractive features of the monograph include an extensive discussion of alternative conjoint models (e.g., additive, multiplicative, distributive, dual-distributive) and procedures for comparing and testing such models. Considerable emphasis is also appropriately placed on the experimental design requirements for the tests. In addition, the author makes clear that commonly used designs (e.g., fractional factorials) typically confound main and interaction effects. I believe it is very useful to make researchers aware of the implicit assumption that interaction effects are zero. This assumption is, of course, crucial to researchers' ability to interpret the results from a "main-effects" design. I find the discussion of testing procedures also very helpful. The commercial practice appears to be to assume a simple model (e.g., part worths) and to use a design that allows for parameter estimation, often at the level of the individual respondent. Market segment or aggregate results are summarized for interpretation. Also, EDITOR: Donald E. Stem, Jr.