Modern Statistics for the Life Sciences Book

This book contains the best explanation of general linear models I could find. Moreover, it is somewhat unusual in that it actually starts with the general linear model, before anything else: before probabilities, before t-tests. It start with attempts to explain variance of data, and proceeds from there. Which turns out to be a very productive approach to the material.

Its main drawback, at least for me, is that it doesn't use R, but sticks to old packages like Minitab and SSPS. This is a bit annoying.

Also it has some quirks; for example every now and then it attempts a so called "geometric explanation of the analysis of variance", in which mean sums of squares are illustrated by Pythagorean theorems in a multidimensional space. I am still to find a person who could understand these explanations, and to whom they would actually help. But they are not long, nor are they frequent, and as long as you ignore them, and as long as you find translating the code to R yourself, it's a truly beautiful book.