DATA ANALYSIS CONSIDERATIONS

The conceptual and methodological framework of environmental studies is multilevel. Conceptually, factors that are believed to effect physical activity behavior are viewed as simultaneously operating at the level of individuals and at the level of neighborhoods. Methodologically, these studies adopt a stratified multistage sampling design in which, first, neighborhoods that match specific characteristics are selected and, subsequently, residents are randomly selected from these neighborhoods. The multilevel structure of the data deriving from a multistage sampling strategy (individual data nested within neighborhoods) call for the use of multilevel methods of data analysis that account for the dependency of observations coming from the same neighborhood. Failing to recognize the dependent nature of the data along with the source of dependency (neighborhood) can lead to finding significant relationships where none exist. From a conceptual viewpoint, multilev el methods of data analysis allow the simultaneous analysis of individual-level relationships (e.g., what is the relationship between age and walking for transport?), neighborhood-level relationships (e.g., what is the relationship between characteristics of the built environment and socio-economic status of a residential area?) and individual-neighborhood cross-level relationships (e.g., what is the relationship between residential density and walking for transport?).

Some good introductory readings about the fundamental ideas underlying multilevel models are:

Bingenheimer JB and Raudenbush SW (2004). Statistical and substantive inferences in public health: issues in the application of multilevel models. Annual Review of Public Health 25: 55-77.

Bullen N, Jones K and Duncan C (1997). Modelling complexity: analysing between individual and between-place variation – A multilevel tutorial. Environment and Planning A 29(4): 585-609.

Hox J (2002). Multilevel Analysis: Techniques and Applications. Mahwah, NJ: Lawrence Erlbaum Associates.

Snijders T and Bosker R (1999). Multilevel Analysis: An Introduction to Basic and Advanced Multlevel Modeling. London: Sage Publications.

Subramanian SV, Jones K and Duncan C (2003). Multilevel methods for public health research. In Kawachi I and Berkman LF, eds: Neighborhoods and Health. New York: Oxford University Press, pp. 65-111.