Ebook: Statistical Inference

Pretty solid textbook overall.

Pros:
1. Thorough introduction to classical statistics. This text hits most of the major topics in frequentist statistics.
2. Clear exposition, at least in the first 7 chapters.
3. Good selection and variety of problems. See below.

Cons:
1. Some important topics are omitted. One that comes to mind is the Multivariate Normal Distribution. A little more Bayesian material would have been nice. Also, some other topics could very well have been omitted which are fairly non standard (ie Union Intersection tests.)
2. Chapters 8-10 are not nearly as strong as the previous chapters. I often had to turn to other sources for additional examples and explanations.
3. Some problems are a complete waste of time. This is my biggest gripe with an otherwise great textbook. Too many problems require drawn out, extremely tedious calculations. Others require leaps of intuition which even seasoned statisticians would not make. My advice is to ask someone who has gone through the problems for recommendations. Or, if a problem is taking a very long time, take a look at the first few lines of the solution - you may save yourself a lot of wasted time.