Power Calculations for Each Hypothesis Posted on January 20, 2009 Guillermo J. Prado (bio) explains why a broad statement about power is not sufficient for a grant proposal.

I don't see how a research application can be funded if you don't have sample size determinations and power calculations in your application. And when I review grants, not only do you have to have a broad statement about power and sample size calculations, but I think you have to have power and sample size for each research hypothesis.

So some research hypotheses may involve a binary outcome while others may involve a continuous outcome, for example. And power for those two hypotheses will not be the same. So generally for binary data, you have a lot less power, so you would need a lot more participants than you would for a continuous variable. So in writing research grant applications, you really do have to have power statements for each research hypothesis.