One skill all professional academics need to learn is how to judge the quality of a piece of work. Not just how competently it was performed, but also how important the work is – is it asking important questions, or is it giving us some crucial new information. I’ve been realising over the last few years that I’m not very good at this, when it comes to ecology. I think the problem is that I’m too interested in the methods, and sometimes fail to see why showing that a power law works is important, because I’m distracted by the triviality of the use of simple linear regression. Which is unfortunate for my own work, because I’m firmly in the camp that says that if a linear regression is all that’s needed, then don’t do anything more complicated. No wonder I never get anything into Nature.
Fortunately for me, INTECOL has had a lot of talks about methods rather than ecology. Indeed, I’m skipping a session like that now, to write this post. With such a broad range of things being done in ecology, it is perhaps not a surprise that it’s difficult to see any one thread running through these sessions. There is quite a bit of statistical development, partly as a result of the on-going Bayesian revolution: Drew Purves gave an introduction to Bayesian thought, and several other people discussed how they have used Bayesian methods for their work: Walter Jetz described using hierarchical methods to fill in missing data, and Gavin Thomas described how he used this approach to examine island evolution in birds, and found a tendency for some to re-invade mainland from islands. In another session, both Nick Golding and Joe Chipperfield talked about how they were using Gaussian random fields to do very different things to model species distributions. In the same session, Tamsin Lee used survival models to show that the Tasmanian Tiger is probably extinct, but we’ll have to wait a few years to be sure (I wonder about the Ivory Billed woodpecker, though).
Between the talks, I have been doing journal stuff – talking to people about possible papers (including telling one person that their idea was, in effect, too good for us), and also discussing the journal’s three year plan, which includes some ideas generated from what I’ve seen at the meeting. I have also so far avoided being interviewed by Barb Anderson for a podcast, but she has a strategy of pushing a sonic screwdriver in front of people’s noses which seems to have worked for getting material. This will be turned into a podcast in the future (near future, hopefully).
The one really big methodological advance we have seen here is in asking plenary questions through Twitter. This has limited the people who can participate, but getting on Twitter is the answer. And it has made the questions much shorter – we have avoided the questioner giving long lectures before ending with “do you agree?”
Dr Bob O’Hara
Editor, Methods in Ecology and Evolution