Approaches to analysis of dietary data: relationship between planned analyses and choice of methodology.

Dietary intake cannot be estimated without error and probably never will be. The nature and magnitude of the error depends on both the dietary data collection methodology and the subjects studied. The impact of particular types of error depends on the question being asked and the analytical methodology used to address it. Examples of these phenomena are presented in this review paper. The future lies in improved estimation and understanding of the error terms and in the development of analytical and statistical methods of coping with these error terms rather than with "improvements" in dietary methodology per se.