I get and agree with what you're saying, but not with how I think you're suggesting it be applied. If you have a screwy design, you can't learn anything. Garbag

I get and agree with what you're saying, but not with how I think you're suggesting it be applied. If you have a screwy design, you can't learn anything. Garbage in, garbage out. However, some methodologies are more susceptible to erroneously thinking you can learn something from garbage. When power analyses are more motivated by funding feasibility and 80% conventions than prior belief, decision theory isn't going to help.