Re: Mixed 2-way ANOVA

Thom wrote: 
Sample size depends on the effect size you want to be able to detect. A
pilot study is used to estimate the effect size.* Unless you have
another way to estimate the effect size (e.g., from previous research
or because you know how big an effect with be of practical theoretical
or clinical interest) there is not much point doing a power
calculation.**

If the study involves placing participants at risk or has unpleasant
consequences (relative to normal everday activities) you probably
should abandon it. If not, and if other ethical requirements are met
just collect the biggest sample you can given other constraints.

* though pilot studies tend to overestimate effect size.
** Some people might advise using canned effect sizes such as d = 0.5,
but my personal view is that it is just a waste of time to use an
arbitrary value.

Thom


Here's a short paper that reinforces Thom's point about using canned effect sizes.

  http://www.stat.uiowa.edu/~rlenth/Power/2badHabits.pdf

--
Bruce Weaver
bweaver@xxxxxxxxxxxx
www.angelfire.com/wv/bwhomedir
.