1 Comment
Guest *Frank* @ 2010-04-15 06:43:48 originally posted:
Hi Yihui,
I enjoy reading your blog.
Your description about Chinese economic journals are generally correct, at least a few years ago. I don't know what do these journals look like today though. However, don't despise economics papers. American's are much more rigorous. Take a look at an issue of Econometrica.
Also, you mentioned bootstrap as an automatic procedure. While in my mind we have to do some serious work to justify, as the bottom line, the consistency of the bootstrap method in your applications, except in a few very straightforward scenarios. Mathematical analysis will never be dated.
By the way, I really appreciate your contributions to R.
yihui 2022-12-16 18:51:33Thanks very much for your reply, Frank. I did read a few papers in Econometrica before I came to the US and I was not able to judge their value because I usually found them full of math and beyond my understanding. I do not mean to downgrade math itself (without which statistics cannot survive) -- my point is, do we really need so complicated methods? (see, for instance, http://yihui.name/cn/2009/03/financial-crisis-and-copula/) I don't know. Maybe ``the ivory tower'' exists for a reason.
I agree with you that math analysis can never be dated -- that is similar to my opinion ``logic analysis should never be replaced by statistical modeling''. Extremism is dangerous.
However, math itself is not the destination for statistics. We need to think about the practical problems instead of enjoying the beauty of asymptopia (asymptotics) even if we only have 3 data points at hand. I'll post my opinions on a question like this raised in my guestbook by another reader soon.
As for R, I'm so happy to know my work is not completely ignored by people :-)
Originally posted on 2010-04-15 07:48:48
Sign in to join the discussion
Sign in with GitHub