ÂÜÀòÉç¹ÙÍø

I need to calculate the required cell size in an experiment using the estimated effect size and the required power. ANOVA is used for analyzing data that is generated by a simulation program. I am trying to identify the number of required subjects in each cell using the effect size and required power.

I have checked a few software programs and they only allow for calculating cell size in experiments with up to 2 or 3 interaction effects. I suspect this is due to the limitations in experimental designs related in lab experiments regarding complexities and confounding effects that will happen with higher number of groups and interactions. However, I am using simulation data and so don’t have most of those limitations.

I also went back to Cohen’s book—Statistical power analysis for the behavioral sciences. The tables in this book use degrees of freedom of the numerator in ANOVA to calculate the power (in addition to other factors). However, these tables also only have values for up to 24 degrees of freedom for the numerator. My experiments have higher degrees of freedom. The book does not describe how to calculate the sample size for higher degrees of freedom. I appreciate any help or suggestions regarding how can I calculate the required cell size for achieving a required power for experiments with higher degrees of freedom in the numerator?

Thanks,

Mohammad