Cluster Analysis Identifies Subpopulations for Health Promotion Campaign Design


Michael Mackert, Department of Advertising, The University of Texas at Austin, 1 University Station A1200, Austin, TX. E-mail:


ABSTRACT While health communication campaigns have been effective in addressing a variety of health concerns, even broadly successful campaigns can miss particular subpopulations. The statistical technique of cluster analysis, which makes it possible to group individuals based on sets of identifying variables, is a statistical method that could prove useful in the design of more effective communication campaigns. This paper illustrates the use of cluster analysis to group women based on their (1) prepregnancy weight, (2) weight gain during pregnancy, and (3) weight retention after giving birth as it relates to the process of targeting subpopulations and developing more effective health communication campaigns and interventions. The implications of cluster analysis, from guiding additional formative research to development of health communication strategies, are discussed.