ABSTRACT

It is a capital mistake to theorise before one has data.