Statistical power to detect main and interactive effects on the attributesof small-mammal populations

Statistical power is an important consideration in the design of experiment s, because resources invested in an experiment may be wasted if it is unlik ely to produce statistically significant results when real effects or diffe rences exist. Using data from toxicological experiments on seminatural popu lations of small mammals, we examined the power of statistical tests for ma in and interactive effects. Our objectives were to evaluate the efficacy of actively reducing within-treatment variation in order to increase power an d compare the power provided by several response variables commonly measure d in Population studies. Controlling population size (N) before treatment i ncreased power to detect effects on N but decreased power to detect effects on population growth (r). For a specified reduction in N, r provided highe r power than N. Fractional measures of recruitment generally provided low p ower, especially-when N was low (<20 animals). Power to detect an interacti on of two adverse treatments depended on the magnitudes of their main effec ts, as well as the magnitude of interactive effects. Estimating or predicti ng effect size is more complex and difficult for interactive effects than f or main effects. We conclude that researchers can increase the probability of detecting real effects by choosing response variables with relatively lo w inherent variability However, efforts:to actively reduce within-treatment variation may have unanticipated repercussions in natural systems.