Several different assays for prostate specific antigen have become available for use in clinical laboratories. We compared the 2 most widely used assays, the Tandem-R and the Pros-Check prostate specific antigen assays, to determine if different assays yield comparable results. We analyzed 70 serum specimens from patients with suspected or known prostatic carcinoma by both assays. Results from the assays showed close linear correlation (r equals 0.988) but the Pros-Check assay yielded values 1.85 times those of the Tandem-R test. The proportional bias between assays is owing to differences in the values assigned to calibration standards in the assays, and it demonstrates a need for improved standardization of prostate specific antigen assays. Important consequences of the bias between assays are that different normal ranges apply and that values from the 2 assays cannot be compared directly.