More than 100,000 clinical caries trials were simulated utilizing the negative binomial distribution on a UNIVAC 1100 computer. In each trial, the raw data and logarithmic transformed data were analyzed using a standard t test analysis. The results were summarized as estimates of the power functions for each case. In almost all cases, the transformed data gave greater power than did the untransformed data. In several common situations, the differences were dramatic.
No takes yet. Share an insight, caveat, or question.
Smith et al. (1983) studied this question.
Synapse has enriched one closely related paper. Consider it for comparative context: