What factors make for a great statistics book? In other words, what writing styles, visualizations, explanatory methodologies, text length(!), tone, depth, etc, etc contribute to giving the audience a great education, understanding, or perhaps even entertainment? The question could perhaps also be framed as "What advice would you give to someone writing a stats book?"

I'm not looking for any specific books, but of course, example are welcome if they epitomise a particular quality. Factors that apply to texts other than books are also welcome.

  • 1
    $\begingroup$ I was thinking about writing a similar question about how people evaluate introduction to statistics textbooks. I wonder though if it is too broad to be answerable. Perhaps someone bold enough can find similarities through lists of great stats books (we have one here on this site). Certainly Tufte's book is what first comes to my mind, but I would be at a loss if you forced me to put my finger on why that is so in any cogent way. $\endgroup$ – Andy W Feb 29 '12 at 12:57
  • $\begingroup$ It's not broad - I'm not asking for every factor, just which ever ones stick out. Tufte did come to mind while I wrote this, as did Huff's How to Lie with Statistics. That question is a good complement to this one. This question should help bring out the common qualities of those texts (as well as perhaps point out some of their deficiencies.) $\endgroup$ – naught101 Feb 29 '12 at 13:42
  • $\begingroup$ I was thinking broad in terms of the breadth of subjects that statistics books cover. For instance it is difficult to compare "whats good for an intro stats book" compared to Tufte's work or any smaller in scope but more advanced methodology book. Audience and goals are so different. Hopefully someone more imaginative than me can identify such commonalities though. $\endgroup$ – Andy W Feb 29 '12 at 14:01
  • $\begingroup$ I specifically didn't mention content, for that reason, although you're right about the content affecting the the design decisions. I guess this could be a quite large list, but I'm just looking for single factors, or a short lists, as answers, and then people can vote as they like. If someone wants to specify a factor that's only applicable to a certain area of study, that's ok too. $\endgroup$ – naught101 Mar 1 '12 at 0:43
  • $\begingroup$ "What advice would you give" makes it clear this post seeks opinions. $\endgroup$ – whuber Feb 19 '17 at 14:52

Just a brief point... When I read a stat book, I like when the data sets are made available so that the examples are reproducible.

  • $\begingroup$ True, I hadn't thought of that, but that's a nice point. It's also a big plus for the R help files (another kind of text), since all the examples there use simple data sets that are included with R, or with the library by default. $\endgroup$ – naught101 Feb 29 '12 at 13:36

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