A COMPARISON OF 2 STATISTICAL APPROACHES TO ESTIMATE LONG-TERM EXPOSURE DISTRIBUTIONS FROM SHORT-TERM MEASUREMENTS

Two statistical approaches are briefly reviewed, both of which are sui table for estimating interindividual variation in long-term exposure: a recently published regression approach and the standard ANOVA approa ch. Simulation studies illustrate the performances of the two approach es in estimating the relevant parameters. Their relative advantages an d applicability are discussed. It is concluded that when repeated expo sure measurements from the same individuals are available, ANOVA is pr eferable. The regression approach however has its place because it can be applied to certain data types where ANOVA does not apply.