Sunday 4 September 2011

Internet Heroes: Andrew Gelman

Andrew Gelman is I think the person on the Internet whom I look up to the most, in the senses of being wowed by his cleverness, wanting to emulate how he writes and thinking he is a great human being.

Cleverness: Gelman was a major pioneer of Bayesian statistical methods, writing a fairly definitive textbook on the subject as well as intellectually championing it through a fairly vicious turf war in the 1980s and 90s. He does loads of really important research into stuff that makes a massive difference to lots of people. These include why it can be sensible to vote, the importance of distinguishing small but important effects from random variation (this is a massive one in my opinion, and features a bonus Kanazawa-bash) and how to display information well as well as reams of just as vital technical work on the theory and computer programmes that help people everywhere draw conclusions from data. It is not just the quantity of his cleverness that I think is look-up-to-able, but also the quality: he sees important problems and sets out to solve them, without falling into the trap of rushing to conclusions, ignoring the technical difficulty of actually finding important things out - loads of people in debates about politics and arguably the whole of economics suffer from this - or getting so into the technical stuff that it becomes it's own reward.

Not only is Andrew Gelman super smart, he also seriously stylish when it comes to writing. He has that knack of explaining a subtle point from first principles without it ever seeming complicated. He generally does this by pretending to go through his own thought process in a really funny and self-deprecating child's "Look! I found ten dollars!" kind of voice. Pretending to be simple allows him to explain things in the simplest way possible without seeming patronising, and also frequently lets him get away with delicious ironic caricatures when summarising other peoples' positions. I think writing in this purposefully unsophisticated common sense way must also help to push back against authority and received wisdom. It's such a key thing to understand that "putting it so simply a child could understand it" actually means "putting it so simply a child could say it"

Hearteningly, Gelman works, writes and thinks on style really hard. He recommends the book 'How to Talk so Children will listen and Listen so Children will Talk' as a how-to and often goes into thoughtful detail about his role models (generally Philip Roth, John Updike, Norman Mailer etc.) and the mechanics of their writing. He also gives us this enigmatic aphorism:

"Style is a subset of content, but the converse holds also: content is a subset of style"

Other than the no doubt wonderful effects of his academic and blogospheric work I don't have any direct evidence of Andrew Gelman being a fantastic human being. Maybe I've been duped by his style! On the other hand those things are pretty suggestive and, what's more, I saw him do a talk once and he sounds like John Malkovich.

Anyway the main thing you should do is READ THIS BLOG PLEASE.



No comments:

Post a Comment