We describe an evaluation of four ELISA methods, including three commercial kits, for measuring recombinant and natural human interferon-g (hIFN-g). Using a panel of samples, including well-characterized reference standards, we compared relative quantification between assays, within assays and, where possible, the absolute accuracy of quantification as compared to other analytical methods. The four assays generated markedly different results; up to an almost 60-fold difference between the highest and lowest values for one sample. The differences between assays were not necessarily predictable. No single correction factor could be determined to correct results from one method to another across the panel of samples tested. We conclude that investigators should be diligent to revalidate commercial methods before depending on such methods and resultant data.