It’s Not The Size Of The Data. It’s The Question It Can Answer.

Rarely do you find a statistician who is able to convey, in large part, the complexity of statistics and appropriate use, but then again rarely do you have a statistician who is able to write in an engaging way for the broad general public. In The Art of Statistics: Learning from data, David Spiegelhalter achieves something usually quite unreachable: Educating any smart layperson about different ways of thinking about the current state of data science, statistics, and artificial intelligence, and more broadly about how to think about numerical information. It is a delightful book to read and certainly shows how working with and communicating numerical information is very much an art as well as a science.

Why did you write this book?

There are many new people coming into data science, and using statistical methods without a full training. And since I feel the traditional way of teaching statistics is too rooted in mathematics and inappropriate for most people, I thought there was a need for a book that introduced all the ideas of a first course in statistics, with minimal mathematics or notation. Crucially, this means changing the order of the content so that probability comes late on.

Why should the public care about better understanding statistics?

The pandemic has shown the vital role that statistics play in our lives, and I feel data literacy is a basic skill that is essential to both citizens and professionals.

Has the advent of “big data” improved or deterred public understanding of statistics and science?

I think this term is rather dated: the size of the data is not particularly relevant, far more important is what questions it can answer, how it is handled and the interpretations that are made. Public understanding depends on trustworthy communication of data and claims made on the basis of that data, which is a generic issue and unrelated to the size of the data-base.