Statistical significance of sediment toxicity test results: Threshold values derived by the detectable significance approach

Abstract

A number of methods have been employed to determine the statistical significance of sediment toxicity test results. To allow consistency among comparisons, regardless of among-replicate variability, a protocol-specific approach has been used that considers protocol performance over a large number of comparisons. Ninetieth-percentile minimum significant difference (MSD) values were calculated to determine a critical threshold for statistically significant sample toxicity. Significant toxicity threshold values (as a percentage of laboratory control values) are presented for six species and nine endpoints based on data from as many as 720 stations. These threshold values are useful for interpreting sediment toxicity data from large studies and in eliminating cases where statistical significance is assigned in individual cases because among-replicate variability is small.