This paper is concerned with the application of market research techniques to segment large populations into homogeneous units in order to improve the reach, utilization, and effectiveness of health programs. The paper identifies seven distinctive patterns of health attitudes, social influences, and behaviors using cluster analytic techniques in a sample drawn from four California cities (N=2,504). The paper finds that these lifestyle clusters predict associated behaviors, including seatbelt use, vitamin use, and attention to health information, with higher means for health-oriented clusters in each case. Furthermore, the paper finds that the clusters predict self-reported improvements in health behavior--e.g., eating less salt and losing weight--and self-reported new exercise as measured in a two-year follow-up, with higher means again found, with two exceptions, for health-related clusters. Implications for public health education and intervention efforts, and future research directions are discussed. Four tables of data and eight notes are included, and 17 references are attached. (Author/SR)