Multilevel health promotion research: conceptual and analytical considerations

Health promotion research is often conceptualized through the use of socioecological frameworks. This results in data or variables associated with multiple levels such as individual, community, and provincial. These data are nested, or clustered. In other words, multilevel health promotion research is based on the idea that community influences health, above and beyond one's individual characteristics or behaviours. These contextual effects can be analyzed rigorously using multilevel modelling (MLM), thus determining whether contextual effects are truly derived from context or are the result of residents' social profile. MLM also facilitates examination of cross-level interaction effects. The authors discuss conceptual and methodological issues related to multilevel research. While multilevel pathways to health outcomes have been suggested at the conceptual level, analytical techniques that produce only average overall effects fail to reveal the various other influences on health behaviour.