The statistical power and biological information of two Daphnia juvenile production test designs

Abstract

This paper examines the statistical power for two designs of Daphnia chronic toxicity test, one (the “old” OECD method) using four groups of 10 animals at each treatment, the other (the “new” OECD method) using 10 individually housed animals at each treatment. Both use reproduction as the end point. It was demonstrated during a pilot ring test of the new method, run concurrently with the old method, that their statistical powers were similar. The old method had a power for detecting a 20% real difference of 0.97, while the power of the new method was 0.90. These findings are reported, together with observations on the quality of data that may be obtained using the different test designs. It is suggested that the new design may be of use in investigating toxicological modes of action, in view of the improved data capture that it allows.