Edinburgh Castle
Head First Statistics
AuthorsDawn Griffiths
PublisherO'Reilly
ISBN978-0-596-52758-7
DateAugust 2008
Pages677
Price£21.99
ReviewerRoger Spooner
Cover image for Head First Stats

This is a very readable study book for new university students (or people around that stage) learning statistics and probability for an academic course. It goes to great lengths to be fun and memorable to read for people who probably have a lot of books to get through, and who may find the topic difficult. It uses clip art and photography extensively, as well as breaking up the routine of the text with different styles of content, and lots of exercises. A typical page has a simple graph or algebraic equation with "hand written" annotations emphasising key characteristics, then a photograph of a student (to whom the reader can relate) making some comment, and a few paragraphs of text. It adds up to something covering many pages, but which is easier to digest.

The publishers make a big deal of the "Head First" style, beginning with a whole chapter about it. The repetition works quite well although I sometimes think that I'd rather cram something new in to my fading grey-cells than hear the same story again. My favourite part of the technique was the brief annotations on pictures and in the observations by the photographed people. They often summarised a tricky point very clearly and concisely. Making a formula talk about itself and who it is similar to, is surprisingly good. Sometimes the repetition got the better of me, for example on pages 416, 417 and 418 where a point that was merely about an example was made 4 times.

There are lots of really well designed examples. They are based on plausible scenarios such as a gym that wants to allocate new members to the class with the closest average age, but keeps getting into a muddle because they do their statistics wrongly. In one, a 17 year old man joined a class where the median age was exactly 17, only to find out that it was a mother-and-toddler swimming class and he was not welcome! The exercises build on these examples. Although I found them generally easy, I think that they would be well balanced for a person meeting these topics for the first time.

For myself, I was looking for a reminder about which statistical techniques can be applied to different types of experimental data, but I did not find that; the pinnacle of the book's coverage seemed to be the normal distribution. I was nevertheless pleased to find a page about sources of bias in badly chosen sample sets. Its main topics go from plotting simple data to averages and variance, an introduction to probabilities, the binomial, geometric and poisson distributions, the normal distribution, sampling and making predictions about the larger population, hypothesis tests, then finally the chi-squared test and correlation lines.

There were some over-simplifications, such as that time should always go on the X axis of a line chart, whereas I'd expect to see the controlled variable which isn't necessarily time. Some charts had compressed areas representing large gaps in the data set which is a trick to be used sparingly, and actually worked against the topic at hand. I found a few mistakes, such as a graph axis on p38 labelled "Hours" which should be "Age", and a data set on p61 miscopied from p55. The print quality is not as good as books on artistic topics; most of the photographs are too dark.

In summary, Head First Statistics covers its topic very clearly, and is an excellent study aid for basic probability and statistics. It looks very different from the average book, and almost all of those innovations work really well, especially the one-sentence annotations. The examples and exercises are strong, and the book progresses well through advancing topics.

Contents

Intro
1  Visualizing Information: First Impressions
2  Measuring Central Tendency: The Middle Way
3  Measuring Spread: Power Ranges
4  Calculating Probabilities: Taking Chances
5  Discrete Probability Distributions: Manage Your Expectations
6  Permutations and Combinations: Making Arrangements
7  Geometric, Binomial, and Poisson Distributions: Keeping Things Discrete
8  Normal Distribution: Being Normal
9  Normal Distribution Part II: Beyond Normal
10 Using Statistical Sampling: Taking Samples
11 Estimating Your Population: Making Predictions
12 Constructing Confidence Intervals: Guessing with Confidence
13 Using Hypothesis Tests: Look at the Evidence
14 The Chi Square Distribution: There’s Something Going on
15 Correlation and Regression: What’s My Line?
Appendix  i : Top Ten Things We Didn’t Cover
Appendix ii : Statistics Tables

faded separator

Page maintained via github.com/edinburgh-pm/edinburgh.pm