by James D. Williams, Virginia Polytechnic Institute and State University, Jeffrey B. Birch, and Steven Walfish Statistics don't lie, but if you don't include appropriate data, the resulting statistical analyses can be misleading. In biopharmaceutical development, if the variability from several different microplates is not addressed, assays testing the relative potency against a standard can be inaccurate. A statistical method that accounts for plate-to-plate heterogeneity is needed ? and presented here.