Skip to content
Prev 164223 / 398513 Next

Power calculations where two samples are of unequal size

Hi Farrel,

I usually simulate in cases like this - pick the effect size and 
distributions you conjecture, simulate your data 10,000 times and look 
how often t.test() gets you a significant difference.

Good luck,
Stephan

Farrel Buchinsky schrieb: