The type I error rate for in vivo Comet assay data when the hierarchical structure is disregarded

The Comet assay is a sensitive technique for detection of DNA strand breaks. The experimental design of in vivo Comet assay studies are often hierarchically structured, which should be reWected in the statistical analysis. However, the hierarchical structure sometimes seems to be disregarded, and this imposes considerable impact on the type I error rate.

This study aims to demonstrate the implications that result from disregarding the hierarchical structure. DiUerent combinations of the factor levels as they appear in a literature study give type I error rates up to 0.51 and for all combinations the type I error rate is greater than the nominal _ at 0.05.

Closed-form expressions based on scaled F-distributions using the Welch-Satterthwaite approximation are provided to show how the type I error rate is aUected. With this study we hope to motivate researchers to be more precise regarding the exposition of the statistical methodology and to suitably account for the hierarchical structure of Comet assay data whenever present.