# Statistics library with knapsack constraint

Suppose you had $200 US to build a (very) small library of statistics books. What would your choices be? You may assume free shipping from Amazon, and any freely available texts from the internet are fair game, but assume a 5 cent charge per page to print. (I was inspired by a mailing from Dover books, but most of their offerings seem a bit out of date). ## 6 Answers Statistical Models: Theory and Practice - US$ 40.00

Grand total of US$167.11 As chl suggested, you still have money left to print nearly all of Hastie (~746 pages, or$37).

You might want to spend $1.20 printing out Matthias Vallentin's probability and statistics cheat sheet. • The last page alone is worth$1.20! maybe we can wedge that into @Carlos Accioly's knapsack by buying one of the books used.. Oct 6, 2010 at 17:26
• The last page comes from this article: math.wm.edu/~leemis/2008amstat.pdf. I've posted a condensed version with links for footnotes here: johndcook.com/distribution_chart.html Oct 6, 2010 at 17:28
• The link in the post is 404; John's links in the comments are still there though. (Edit: I have located the new home and updated the original post) Nov 4, 2015 at 1:31
1. Harrell, FE. Regression Modeling Strategies (Springer, 2010, 2nd ed.)
2. Izenman, AJ. Modern Multivariate Statistical Techniques: Regression, Classification, and Manifold Learning (Springer, 2008)

You should have money left to print part of The Handbook of Computational Statistics (Gentle et al., Springer 2004) and The Elements of Statistical Learning (Hastie et al., Springer 2009 2nd ed.) that are circulating on the web. As the latter mostly covers the same topics than Izenman's book (as pointed by @kwak), either may be replaced by one of the Handbook of Statistics published by North-Holland, depending on your field of interests.

• 'Elements of...' largely covers the same topics as 'Modern Multivariate...'. Oct 4, 2010 at 14:30
• @kwak Well, some savings for @shabbychef! I found Izenman's book more an applied textbook featuring some methods not covered in ESL (e.g. correspondence analysis), but you're right they are pretty written in the same spirit one each other. (I'll update my answer)
– chl
Oct 4, 2010 at 15:11
• > i agree with you btw (i own both books). On the other hand, if budget is a binding constraints, you might want to diversify. Oct 4, 2010 at 19:27
• @kwak For sure :) Add your must-have references! (thx for the undelete)
– chl
Oct 4, 2010 at 19:58

As a social scientist I would have to vouch for the Sage Green Books. If you are a bargain shopper you would be able to get between 10 to 20 books for 200 dollars (assuming no shipping). For those not familiar these are all introductions to various methodologies aimed at people with no more than an introduction to statistical inference in most circumstances.

What topics are you interested in? I learned from KNNL ($157.50), but oh gosh I couldn't imagine carrying this thing around -- you'd be asking for a reading list on scoliosis correction. "General Statistics" is certainly an area of interest, but I'm curious if you're more interested in depth, breadth, or some mix of both. A tad pricey, but Bruning and Kintz's Computational Handbook of Statistics ($95.80) would certainly fit in your knapsack.