I have been working on a spreadsheet for sample-size estimation for some time
now. I have uploaded a draft form of it to the files section of this list.
Download it by going to
http://sports.groups.yahoo.com/group/sportscience/files/ and clicking on the
obvious file.

The spreadsheet has a new way of estimating sample size based on clinical
decisiveness, in addition to the traditional method based on power for
statistical significance. Most outcome statistics and straighforward designs
are covered, including repeated-measures for numeric variables, but not repeated
measures for nominal variables. Comments welcome.

Will