This study was undertaken to compare the sensitivity of two in vitro screening test methods and to determine the accuracy of predicted response to spiked laboratory water samples. A newly developed enzyme-linked receptor assay (ELRA) and a widely used yeast estrogen screen (YES) assay were selected to evaluate the estrogenic responses. Four natural, pharmaceutical, xenobiotic or phytobiotic chemicals: 17beta-estradiol (E2), tamoxifen, bisphenol-A and resveratrol were examined, and 17beta-E2 was used as a positive control. 17beta-E2 can strongly induce estrogenic response in both test systems, however, ELRA was found to be more sensitive to 17beta-E2 with a detection limit of 0.07 microg/l compared to 0.88 microg/l in YES assay. Similar results were obtained for bisphenol-A and resveratrol, and their estrogen potencies relative to E2 (100%) determined by ELRA were at least 5.6 times greater than produced by YES assay. ELRA was unable to distinguish the anti-estrogen tamoxifen and YES assay is also poor at distinguishing. Comparison of response to spiked laboratory water samples show that ELRA can give accurate determination to all four chemicals with recoveries among 70-120%, while YES can only give accurate determination to 17beta-E2 and bisphenol-A with recoveries among 69-112%. The comparative results provide evidence that ELRA is more suitable for rapid screening estrogenic potency of the environmental samples. Combination of ELRA and mammalian cellular assay will constitute an advantageous test to specify agonistic or antagonistic effects.