Top critical review
352 people found this helpful
on November 27, 2007
I have a lot of mixed feelings about this book.
As a graphic designer and a minimalist, I love the way this book looks and I love the graphics Tufte's team has created.
Yet, the minimalist in me also dislikes Tufte's prose, which is surprisingly un-minimalist. The text is repetitive, and although Tufte does use this effectively at times to reiterate or summarize concepts, there are far more instances where I feel the repetition is simply irritating (Tufte's poems and block-quote summaries are, to me, good examples of this).
The minimalist in me is also not fond of the nature in which Tufte presents his opinions. Tufte makes frequent use of words like "lies" and "tricks," and while I am not fond of the targets of Tufte's derision, I feel that use of these words unnecessarily and unfairly assumes that poor graphs are always the result of malicious intent. Tufte's presentation as a whole, I feel, is often unnecessarily condescending (see e.g., p 120); indeed, Tufte seems to feel that unenlightened minds somehow deserve our ridicule and contempt.
As an academically oriented statistician, I also have mixed feelings. I give Tufte an immense amount of credit for opening a dialog about statistical graphics. And, I am grateful to him for pointing out the flaws and "wrongs" in the ways in which statistics are so often presented and suggesting ways in which these approaches can be changed. Moreover, I happen to agree tremendously with a large amount of what Tufte has to say, and often passionately so.
That said, I am puzzled by the amount of relevant concepts which are omitted from this text (or merely brushed over). Good examples include: samples versus populations, continuous versus categorical data, and exploratory graphics versus graphics presented for presentation.
For that reason, the academic and statistician in me is watchful of Tufte's role as an instructor of statistical ideas. Much of what Tufte has to say is not in fact unique or necessarily "right," and also not nearly close to being all there is to be said about statistical graphics (even at an introductory level). If students allow this text to be the sole contribution to their statistical education, I fear that -- without statistical intuition or knowledge to draw from -- they will not be critical statistical thinkers but blind followers. (Of course, none of this is intended to be a criticism of Tufte or Tufte's book.)
Those seeking a good overview of statistical graphics: keep in mind that this not strictly an instructional book. And while I wouldn't discourage you from reading or buying this text, I also wouldn't discourage you from seeking additional resources, either as an alternative or a supplement to Tufte's works. Much of the ideas supplied by Tufte here -- plus a great deal more -- can fundamentally be found in a good introductory statistical course or text, either directly or indirectly. Moreover, I would argue that there is absolutely no substitution for such an education.