The Centre for Open Science recently set up the Pre-registration prize – $1000 for 1000 researchers who preregister the designs and aims of their studies (see the nice video about it here). The aim to make research more transparent and allow us to identify studies that set out with some hypotheses, test them and publish the results and those studies that start with hypotheses, test them, find they don’t work and then decide to test something else instead. The pressure on researchers to publish significant results is so high that often that we often carry out exploratory analysis, see what relationships look like and then test them, writing up the paper as if they tested some a priori hypotheses. Maybe pre-registration can help to battle this. Ultimately though, for this idea to work journals would need to demand registration of study designs prior to publication, and I think we are some way off that right now.
Nature just published a nice guide on how to we could do a better job of plotting complex data. Given that I have often resorted to converting things into proportions so that everything is more comparable, this is an approach I will be using more often if I can. (HT statsforbios)
A recent post by Kent Anderson of Science on the Scholarly Kitchen blog has caused a bit of a rumpus. I’m not saying what side I come down on, I’ll just leave this here and let you make up your own mind. (HT @recology_)
I discovered from this Buzzfeed post that the annual ESA meeting and the world’s biggest my little pony convention (apparently called bronycon) will be in the same convention centre… at the same time. I’m not going, but really wish I was, if only because ‘my little pony’ fans look set to outnumber ecologists about 2:1. (HT to @markus_eichhorn).