Although excellent overall, the VSI series is variable in quality, with some gems (Psychology, Anthropology, et al) and some dross (Architecture, Locke, et al). This particular entry is somewhere in the middle. It starts very well but gets bogged down somewhat toward the end.
A problem that runs through the series is the inordinate number of typos - the proofreading in the earlier editions was very poor. OUP seems to have fixed this, and I am happy to say I saw no errors in this book.
But another recurrent problem with the series - poor use of illustrations - does afflict this book, except that here it is simply the paucity of illustrations that is the problem. Statistics is a very visual subject and there are ideas which are very simple when presented in graphical form but extremely difficult otherwise. In the second half of this book, from chapter 4 onwards, when things get technical, there is a need for a great many illustrations, almost on every page. In fact, there are only eight illustrations in the entire book. Ironically, three of those are in a brief section called Statistical Graphics and do not relate to anything described in the text but are used simply to illustrate how helpful graphics can be! Hence, the second half of the book is much heavier going than it need be.
There is however much that is good in the book. There is a useful emphasis on the difficulty and importance of collecting good quality data. This, and a reference to the Sally Clark case, are among the ways the author grounds the subject in reality. He presents a good case for statistics being at the very heart of a great many disciplines.
So, a good short introduction, spoiled by having too few illustrations.
[PeterReeve]