The purpose of this paper is to describe a method for determining whether a particular procedure for estimating regional metabolism using the deoxyglucose tracer analogue yields better data than another in terms of subsequent statistical analysis. The method is based on a simple model of regional cerebral glucose metabolism with three potential sources of metabolic variability, namely individual differences in cerebral metabolic rate, consistent regional differences and error. When the literature rate constants were compared to a dynamic procedure for estimating regional rate constants in patients with Huntington's Disease, the literature values were clearly superior in that the error component was approximately half (18.5 versus 39.3%). Although these results cannot be generalized to all procedures for estimating regional glucose metabolism, the method can be applied to determine if a particular procedure will be more sensitive than another to differences between groups.