Diversity frequently is defined for policy and research purposes by broad social demographic categories such as race, ethnicity, and socioeconomic status as well as gender and age. This has some advantages. These characteristics are apparent and are often easily measured, and thus comparisons between groups are facilitated. Also, for many health behaviors and health outcomes, researchers using these categories have located important disparities. This evidence has allowed groups with poor health outcomes and their advocates to make a strong moral claim for redress and reallocation of resources. Also, groups organized around these identities have forged political bonds, and that has given them some power and ability to make additional claims on resources.

At the same time, these diversity categories may have more political relevance than substantive relevance in the construction of communication programs. There is often as much heterogeneity with regard to a behavior and its determinants within a specified group as between groups. Also, the use of such categories to report disparities may reinforce stigma, a sense that a disadvantaged group is incapable of helping itself. This may be a particular concern when race and ethnicity are used as the criterion for diversity.

There are alternative ways to describe heterogeneity in the population that do not rely on these categories.