What are some good graduate level statistics textbooks?

[kuahji]

Did a search on statistical textbooks on the forum but did not find much except introductory statistical textbooks. Just curious as to what are some of the better intermediate to advanced level statistics textbooks on the market?

 

[chiro]

Hey kuahji.

The one I am using is this which includes a lot of "beginners" stuff:

https://www.amazon.com/dp/0534377416/?tag=pfamazon01-20

To be perfectly forthright, I used mostly a lot of lecture notes provided specifically by the university to learn a lot of the statistics and the book is mainly used as a reference if I needed a derivation or wanted to understand some method or terminology.

It has the same sort of problems that I did in the university course (and some of the problems were directly from the book), but again a lot of the learning came from a variety of sources and not just the one book.

One book for applied probability that I found useful is this one:

https://www.amazon.com/dp/0125980620/?tag=pfamazon01-20

If you go into advanced statistics you will at some point need a book on Generalized Linear Models and the Bayesian framework of Statistics.

I haven't used any books on both of these so I can't recommend any for the GLM's or Bayesian models (I used mostly lecture notes for these).

 

[jasonRF]

Common intermediate level statistics books include, "Introduction to mathematical statistics" by hogg and craig, and "statistical inference" by Casella and Berger. I own the 1st edition of Casella and the 5th edition of Hogg and Craig. Hogg and Craig is perhaps upper level undergrad (so may be classified as intro by some folks), while a quick google search tells me that Casella seems to be used very often for intro grad courses in a variety of departments (math, statistics, economics, operations research, etc). I'm not sure that necessarily means it is one of the better textbooks...

I personally like Hogg and Craig much better, as it seems to have better discussion to motivate why you are doing what you are doing. Also, at least the 1st edition of Casella is essentially univariate only when it comes to estimation, bounds, etc., which is quite limiting. They are both reasonable books, but neither of them are anything special so i would only buy inexpensive used copies of old editions if you are looking to buy. Check out of a library if you can! I would hope true statisticians would jump in here to give their advice (I am an engineer that uses a lot of statistics - for engineers I recommend the statistical signal processing books by Kay).

The most commonly cited advanced books are those by Lehmann. Not sure how good they are...