I am looking at following books:

  • Statistical Models (Cambridge Series in Statistical and Probabilistic Mathematics) - A. C. Davison
  • Essentials of Statistical Inference (Cambridge Series in Statistical and Probabilistic Mathematics) - G. A. Young
  • Asymptotic Statistics (Cambridge Series in Statistical and Probabilistic Mathematics) - A. W. van der Vaart.

They all cover different topics, I skimmed through the pages of both of them and it seems to me they are all useful for data analysis, but I wonder which one would be the most useful to buy / read first?

  • $\begingroup$ Hi there, exactly what type of experimental data analysis do you need? You may require a specialist book. $\endgroup$
    – Michelle
    Commented Feb 11, 2012 at 10:33
  • $\begingroup$ Any reason you're looking at books from only one publisher? $\endgroup$
    – onestop
    Commented Feb 11, 2012 at 13:27
  • $\begingroup$ I own the last two, but not the first. Neither of the last two could be characterized as geared toward data analysis. These are books on mathematical statistics. The van der Vaart text is a beautiful one, but may be rough going for those without sufficient math preparation. The Young & Smith text is essentially a very high-level rapid introduction to some of the core ideas of mathematical statistics. It is a nice read, but lacks much depth in my opinion, and, I think that is intended. The Davison text appears to have a similar intent. None of these are about data analysis, per se. $\endgroup$
    – cardinal
    Commented Feb 12, 2012 at 16:31
  • $\begingroup$ In other words, you might consider editing your question to be more specific about what you are trying to learn. Understanding your goals better will invite more specific and useful answers. :) $\endgroup$
    – cardinal
    Commented Feb 12, 2012 at 16:31
  • $\begingroup$ @onestop Yes they are the only available in my library. $\endgroup$ Commented Aug 8, 2012 at 22:58

1 Answer 1


I haven't read any of those textbooks but a quick look at their table of contents makes me think you might find the first (Statistical Models by A.C. Davison) most helpful. That one takes the most applied perspective; the other two focus on the rigorous mathematical underpinnings of statistical inference. For doing actual data analysis, you will likely be better off with a textbook that has an applied perspective.

Even the Davison text, though, isn't specific to experimental data analysis. Its section on analysis of variance (a key statistical method in analysis of many experiments) is only eight pages long. The chapter on designed experiments is 47 pages long, which may seem like a lot until you realize that entire textbooks are written on design and analysis of experiments. Once you've covered basic statistical ground as in the Davison textbook (or more rigorously in one of the other books), you might be ready to focus specifically on experiments.

Most textbooks that cover data analysis for experimental data also cover the design of experiments. As you can see from questions like this one -- How to test hypothesis for group differences -- the analysis of experimental data is tied up with the design of the experiment. Poor experimental design can ruin your chances of answering the questions you want to answer. Even if you are not the one doing the experimental design, you need to know what a well-designed experiment looks like. Ideally, you will be involved in the design so that you can ensure that the data you'll have for analysis will be useful.

If you do decide to get a book focused on experimental design and analysis, you'll want to get one that is specific to your domain. I like Myers, Well, and Lorch's Research Design and Statistical Analysis. I've found it useful in a social science research setting. But if you're working in another area (clinical trials or ecology, for example), you can find books that cover experimental design and data analysis in those contexts.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.