@gung I would like to ask you whether you have found a solution to your problem explained in stats.stackexchange.com/questions/18119/…? Namely, I have exactly the same issue. I appreciate Harrell's answer to your question, but I am not sure how good my model is, I am not really happy by just having small p values from various tests. Thanks!

No, I never really did, @AjdinHalilovic. The answer seems to be that you shouldn't be trying for that.

That's too bad...

I meant to say large p values :) Thank you for your answer.

It think one should speak about 'family' if one means 'family', and use 'distribution' if one means 'distribution'. Some expression like N(mu,sigma^2) can mean both. Depending on how one views those indexes. However in the example from @gung about wikipedia's gamma distribution, I believe it is wrong. Any variation of the format 'the distribution ... is a family', would be wrong.

Hi, @MartijnWeterings. I'm not sure I follow you here. I haven't mentioned the Gamma distribution or Wikipedia.

sorry, wrong quote (I normally take dozens of edits, not good for a chatbox, and this is my first message here). I meant to quote amoeba last wednesday 23:05

No problem.