The test of significant toxicity (TST) is a hypothesis-testing approach based on bioequivalence developed by the U.S. Environmental Protection Agency (U.S. EPA) for analyzing whole-effluent toxicity (WET) and ambient toxicity data. The present study compares results of acute and chronic toxicity tests of effluent, storm-water, and ambient (i.e., receiving-water) samples using both the TST and the standard no-observed-effect concentration (NOEC) approach. Valid WET data were analyzed from 890 tests provided by more than 25 dischargers in California and Washington, USA, representing the majority of test methods used in the U.S. WET program. An additional 3,201 freshwater chronic toxicity tests, obtained from ambient monitoring programs in California, were also analyzed. The TST and NOEC approaches both declared a low number (<6.5%) of tests toxic if effects were below the unacceptable toxicity regulatory management decision (RMD) of 25% effect in chronic tests or 20% effect in acute tests. However, those test methods having generally lower within-test variability and greater test power (e.g., urchin fertilization test) had a much lower percentage of tests declared toxic than the NOEC approach when effects were below the unacceptable toxicity RMD. In addition, the TST showed fewer tests to be nontoxic than NOEC if the test exhibited effects greater than the toxicity RMD (0.1 and 9.6% for TST and NOEC, respectively, for effluents and 0 and 9.5%, respectively, for ambient samples). Our results demonstrate that the TST is more likely to identify a toxic sample when effects are fairly substantial (≥ 25% effect in chronic testing and ≥ 20% effect in acute tests) and less likely to identify a sample as toxic when effects are negligible (≤ 10% effect). Furthermore, these results demonstrate that appropriate WET data interpretation benefits from having well-designed test methods with sufficient power to identify significant toxicity or biologically insignificant effects when they occur.