Using the sample range as a basis for calculating sample size in power calculations

Lacking a sample standard deviation to use as an estimate of sigma, in samp le size computations, consultants often divide a sample range (R) by a cons tant to estimate sigma. To avoid underpowered studies, the estimate must ha ve a high probability of being greater than or equal to cr. The probability of being, CF greater than or equal to or is estimated for R/6, R/4, and R/ (standardized mean range) for various parent distributions and a range of s ample sizes. Those probabilities are found to be quite low for many distrib utions and sample sizes. Alternative divisors (K) are found for which the p robability of R/K being greater than or equal to sigma is at least 95%.