Monthly Archives: February 2016

Poor review

Firstly, let’s namecheck the wittiest punner in stats for that title, Stephen Senn.

This recent post on Andrew Gelman’s blog is essential reading. I suspect my readers are all over there too, but I’ll mention it here because of this point in wrapping up:

Peer review can serve some useful purposes. But to the extent the reviewers are actually peers of the authors, they can easily have the same blind spots. I think outside review can serve a useful purpose as well.

I’ve seen this a lot in my life as a medical (read ‘health and social care more broadly, with a dash of education’) statistician. There are distinct tribes of healthcare professionals and they do things, including research designs, analytical methods and communicating findings, in their own sweet way. There’s generally no reason, it’s just custom and ritual. If you don’t fit that mould to some extent, you get rejected. Often, I find myself consciously peppering the paper / slides with some shibboleths that will ease my journey to REFland. (Of which, sample size calculations for anything that isn’t a randomised controlled trial is the most common, although I am no stranger to the Totally Unnecessary Reporting Diagram (TURD). I draw the line at Cohen’s D though; D stands for d’oh.)

‘Outside review’ reminds me of the idea of ‘strong inference’, or having your worst enemy analyse your data too and see if they can destroy your conclusions. You don’t have to go that far though, you could just make sure that reviewers extend beyond the specialism and profession of the authors to break that parochialism and question the unquestionable.

People Against Goodness And Normalcy

Essentially, if they can’t understand it, then it’s not written well. I don’t accept any argument that the subject is just too complex for outsiders – because the authors’ interests were once upon a time confined to Lego or the Smurfs, so it must be possible – nor do I claim to have got it right myself – it’s a constant challenge to pitch Bayesian latent variable models just so for a subject-expert audience.

I don’t know about you, but I enjoy reading academic papers outside my own field (OK, no critical theory please, but I’ll consider pretty much anything else). Maybe I should start an occasional series of randomly selected academic papers here, or maybe I just don’t have time for that.


Filed under Uncategorized

Notes after the death of Pierre Boulez

I’m going to take a diversion from the staple statistical fare and mark the passing of a man who has obliquely, and not without contradiction, been a long-running source of inspiration to me. The death of composer and conductor Pierre Boulez was announced in January. There is plenty you can read about him online, so I won’t attempt any kind of obituary; rather, I want to reflect on the art-science intercourse and the unexpected lessons in living and working.

For this post, it was quite hard to decide what to cover and how to structure it. Finally, I felt I should get on with it and follow his style and form. It would be tempting to keep deleting and rewriting it every ten years or so, but I don’t plan to do that. The text is not here but on my website (click here), because it uses a little JavaScript to play the role of a conductor interlocking the rings of Le Marteau Sans MaĆ®tre.

Leave a comment

Filed under Uncategorized