Jordan Lin wrote:
Can anyone give me some guidance as to how to calculate the required sample
size for the following two-way within-subjects design?
Planned parameters of the experiment are:
(1) two independent variables, one with three levels and the other with six
levels,
(2) have each subject give two dependent measures of the same kind to two
different stimuli, e.g., does he like condition A first and does he like
condition B second; each condition is a combination of the two independent
variables,
(3) the dependent variable is a discrete variable (yes/no)
(4) pair-wise two-tailed tests of proportions
(5) alpha=.05, beta=.2, correlation between responses to any two stimuli
ranges from 0.1 to 0.3
--------------------------------------------------------------------------------
Have you considered simulation? See A. H. Feiveson, Power by simulation. _The
Stata Journal_ 2(2):107-24, 2002. The only wrinkle would be to generate
correlated binomial random variates with specified correlation coefficients.
For this, you could use trial-and-error from latent variables generated with
-drawnorm-, or use the method described in A. D. Lunn and S. J. Davies, A note
on generating correlated binary variables. _Biometrika_ 85(2):487-90, 1998.
Joseph Coveney
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
