SHOULD AGGREGATE PRIOR TO ESTIMATION BE THE RULE?

On the basis of a variety of evidence, including the paper with Professor Harold Watts and a paper by the author, the case for seeking and frequently using disaggregated data seemed strong, but one nagging concern remained. Suppose, as seems likely, the microcomponents exhibit different behaviors. In this case it might not be sensible to pool the data and treat it as a single sample from a single universe. This paper copes directly with this issue, and demonstrates the importance of using disaggregated data even when microcomponents. Instead of using a simulation to study aggregation across components with different behaviors, one could have attempted a mathematical deduction of the effects. In order to do this certain simplifying assumptions would need to be introduced in order to make the problem tractable. Another approach to the problem of aggregation is to use data from the real world to fit models at different levels of aggregation, and then to evaluate the relative worth of these models by means of some quantitative measure.