what if i can't decide on an $\alpha$ and just report the p-value without deciding if i can or can't reject the null? i still show that depending on the $\alpha$ what the decision would be, right? i know it breaks a bit from convention but is that so bad?