You are not currently logged in.
Access your personal account or get JSTOR access through your library or other institution:
A Consonance Criterion for Choosing Sample Size
W. Alan Nicewander and James M. Price
The American Statistician
Vol. 51, No. 4 (Nov., 1997), pp. 311-317
Stable URL: http://www.jstor.org/stable/2685897
Page Count: 7
Preview not available
Sample sizes determined by the proposed criterion ensure that subjectively "important" parameter estimates will be statistically significant; alternatively, subjectively "trivial" values will be excluded from confidence intervals for those parameters. Formulas are given for simple mean differences, multiple correlations, and mean comparison methods in multiple group settings, with suggestions for extension to larger, factorial designs. Although no assumptions are necessary about the true values of unknown parameters or a desired level of power, adequate power is provided for important results. Only central distributions are required for either the exact iterative solutions given or their closed form approximations.
The American Statistician © 1997 American Statistical Association