In defense of proper cosinor analysis.

Abstract

Parallaxes can originate from different sources, from bias towards proving a favored hypothesis, but also from bias against existing evidence in support of an unwanted hypothesis. One approach to avoiding either bias is to shift emphasis from hypothesis testing to parameter estimation. Complementary statistical approaches applied to replicates of given experiments can then serve to further a given field of science by providing converging or diverging results depending on whether the original hypothesis was true or false.