I just finished reading Deirdre McCloskey's latest book, Bourgeois Dignity: Why Economics Can't Explain the Modern World. As I'm reviewing it for the New Statesman in the New Year, I won't do so here, but it prompted me to pick up one of her earlier books, How to Be Human – Though An Economist. It's a good-humoured guide to young economists, and indeed older ones, full of advice about to avoid the reductionism and narrow-mindedness that does disfigure so much economics.

McCloskey is in a good position to give such advice, having started as the driest of neoclassical economists and done the maths and heavy-duty econometrics. She now presents herself as a dissident economist, possibly even heterodox (although I think she's still far too rigorous for the many in the heterodox community to welcome). I regard her more as a humane economist, who reminds us how rich a training in economics used to be – compulsory economic history as well as compulsory general equilibrium theory and econometrics, even a foreign language requirement in the distant past.

The essential chapters in 'How To…' are those on econometrics and the shibboleth of statistical significance – distinct, she reminds us, from meaningfulness. This section should be compulsory reading for students. McCloskey has continued to work on statistical con-trickery by economists, with Stephen Ziliak. They recently co-authored The Cult of Statistical Significance, and Ziliak was interviewed on BBC Radio 4's marvellous More or Less.

Having dug the book off the shelves, I might leave it out in a prominent place in case son 2 (a second year undergraduate) decides to pick it up over the vacation. Highly recommended – it's a very nice read and besides I mostly agree with it. And a link to my review of Bourgeois Dignity will follow in due course.