A variety of quality control (QC) analyses are performed with regular frequency in environmental chemical measurements. 'Typically, an acceptable result must be obtained for each QC check before measurement of environmental samples may begin, and each check is periodically repeated to validate the analyses of intervening samples. Corrective actions are performed if the check is failed, and the intervening batch of samples are remeasured. Under such a scheme, excessive costs will occur if the quality control checks are too frequent (i.e., batch size is too small) or i&equent (i.e., batch size is too large). This paper describes an approach for investigating the frequency of QC checks for minimixing the expected cost of measuring a set of samples. Two specific expected cost models are discussed. It is shown that the empirical estimate of measurement drift can be used along with computer simulation to estimate the batch size that will minimize the expected cost of sample analyses. Measurement drift of inductively coupled plasma mass spectrometry for 121Sb is demonstrated with data from standard reference materials repeatedly analyzed between QC solutions. Examples are presented using computer simulation and empirical estimates of linear drift functions for analytical measurements.