One-way fixed-effects analysis of variance
You can use this dialog to solve power and sample-size
problems for a one-way fixed-effects analysis of variance.
Enter information in the windows and click on one of the
three buttons depending on what you want to solve for.
Suggestion: Enable the Java console ("Options" menu)
so that you see any error messages generated by the statistical
routines.
Please let me know if you have trouble or obtain
erroneous results. (Describe what
happens, and tell me about your system, browser,
and whether you've had
success running java applets from other sources).
Explanation of parameters
This dialog is based on the standard assumptions for the
a completely randomized onew-way layout. Independent
random samples
are drawn from a specified number (# of groups) of
populations, each being normal with the same standard deviation
(Within-group SD) but possibly different means.
The sample SD of these means is given by Between-group SD.
We assume that the samples are of equal size n per group.
Alpha is the significance level of the anova F test,
and Power is the probability of rejecting the null hypothesis
of equal means. In practice, pilot data or past experience
are used to obtain an estimate of Within-group SD.
Feedback
is always welcome.
I will be working on more things like this when I can.
Russell V. Lenth (rlenth@stat.uiowa.edu),
Department of Statistics and Actuarial Science,
The University of Iowa, Iowa City, IA 52245