who have an interest in the outcome. Two factors are of particular importance for these decisions: (1) the desires of the intended populations and community leaders, and (2) the projected costs and benefits associated with intervening or not intervening for a particular group.

Consideration of the desires of intended populations requires the involvement of potential audiences in the design of health communication efforts. Campaign planners and implementers must have some measure of understanding, dialogue, and mutually accepted and shared practices with intended audiences (Gbadegesin, 1998) to ensure that the communication campaign addresses shared values and goals (rather than those imposed by implementers). Challenges in identifying relevant and representative group members and social authorities, as well as in determining the degree of consent required from a group prior to the initiation of health communication interventions, are explored in our discussion about the ethics of health communication in Chapter 7.

The majority of health communication campaigns reviewed by this committee selected intended target populations based on need. Need was generally identified through available epidemiological and public health data sources, which characterize audiences by demographic variables (e.g., ethnicity, gender, age, educational attainment, socioeconomic status). In some cases, campaigns relied on other quantitative data for the selection of audience segments, such as survey research assessing knowledge or attitudinal, psychological, psychosocial, or behavioral characteristics of audiences, in addition to the demographic and epidemiological data available from public health sources (see, e.g., Grelen, 2001).

For example, the National Air Bag and Seatbelt Safety Campaign used available data to identify its primary audiences based on populations at highest risk for morbidity and mortality associated with traffic accidents (i.e., first-time parents, young parents, new drivers, new vehicle buyers, and ethnic minorities). Similarly,