Multiple or Repeated Assays

This chapter describes multiple or repeated assays. When an assay procedure is used repeatedly in a given laboratory, an astute investigator is often able to spot results that seem abnormal and should be repeated. To obtain a more reliable estimate of its potency, an unknown preparation may be assayed independently at different times in the same laboratory or in several different laboratories. The problem arises as to how best to combine these data into a single estimate. Two procedures are possible. One is to work directly with the individual log potencies and their errors. The other method is to examine the homogeneity of the standard deviation and slope used in determining each of them and, if consistent, to recompute each individual assay before combining them into a composite value. An assay may be biased by factors that do not increase its standard error. The magnitude of these factors cannot be predicted a priori , but their composite effect is measurable from the agreement of the individual M's.