While obesity is recognized as a major health risk factor, there is a lack of clear and consistent research results as to causes, course, or cure. One source of difficulty is the lack of design and analytic techniques that suitably deal with the facts that: (a) development of obesity, or recovery from obesity, is a process illsuited to study using crosssectional methods; (b) there are individual differences from patient to patient that must be considered in modeling the process; and (c) what determines the process is likely to be multifactorial, and resistant to study with the univariate methods most commonly used. Two simple models, one for early development and one for treatment response, are presented as illustrations of the types of methodologies needed.