Power analysis for the NOEC: What is the probability to detect small toxic effects on three different species using the appropriate standardized test protocols?

Nelly van der Hoeven

Leiden University, EEW, Department of Theoretical Evolutionary Biology, P.O. Box 9516, 2300 RA Leiden, The Netherlands

Present address:

ECOSTAT,
Vondellaan 23,
2332 AA Leiden,
The Netherlands
tel.: (+)31-(0)71-5315011; fax: (+)31-(0)842-116988;
email: NvdH @iecostat.nl

Summary

The power to detect small toxic effects with ecotoxicity tests is investigated on the basis of ring tests for the protocols of three standardized reproduction tests. The test species are Daphnia magna, Eisenia fetida and Folsomia candida, respectively.
The coefficient of variance for reproduction in the controls of these tests is about 14.4% for D. magna, and 19% for both E. fetida and F. candida. Given the standardized test protocols this implies that a reduction inhibition of 20% will be observed in 80% of the D. magna tests and in only 18% of the tests with either E. fetida or F. candida. Comparing 5 different concentrations with one control, 12 replicas are needed per concentration and 27 in the control for D. magna, and 21 and 47, respectively, for both E. fetida and F. candida to observe an effect of 20% in at least 95% of the experiments.
The power analysis is based on a one-sided Dunnett test. This test assumes normal distributed data with equal variance at all treatments. These assumptions are checked for D. magna test using ring test data. The distribution of these data does not deviate substancial from normal. For small toxic effects, no significant correlation between mean and variance of reproduction was observed.

keywords: Dunnett test, Ring test, Daphnia magna, Eisenia fetida, Folsomia candida

Ecotoxicology 7: 355-361, 1998