I have a new paper (with Alastair Wilson) out in the Behavioural Ecology journal, entitled ‘Avoiding the misuse of BLUP in behavioural ecology‘. Our paper is aimed at researchers working on individual variation in behaviour (e.g., personality, behavioural plasticity, behavioural syndromes), particularly those wishing to investigate associations between that behavioural variation and some other trait or variable (e.g., another behaviour, a physiological response, or even some external environmental variable). The thrust of the paper is really a call to ensure that we are using the proper statistical tools to test our hypotheses, rather than using other approaches that are known to give spurious results. The paper is quite brief, and can be found here (or drop me an email if you require a reprint).
Of course, pointing out problems is in itself not hugely useful without solutions being to hand, and so we have provided these in the form of tutorials for multivariate models in the programming language R. We are still working up more tutorials to cover more of the kinds of issues in which people are interested, so do keep checking back for updates – and let me know if there are any other relevant topics that you’d like to see covered! As has been pointed out on twitter, while we focused on animal behaviour because we work in that field, these models are applicable to many other fields in which researchers are interested in the causes and consequences of variation in labile traits.
NOTE
It has been pointed out to me (post-publication) that the Adriaenssens et al. (2016) paper in Physiology & Behavior, ‘Telomere length covaries with personality in a wild brown trout‘, did not use BLUPs extracted from mixed models in a secondary analysis, and is therefore incorrectly included in Table 1 of our Behavioural Ecology paper. I have apologised to Bart for my error, and contacted the publishers to see whether this reference can be removed from the paper.
Despite its great litany of flaws, I love R. I love it super-hard. One of the best things about it is that people are always making weird packages that solve your problems; one of the annoying things is that you tend to find them just after they’d have been really useful, and that maybe they solve only a part of the problem. But it’s the beauty of open-source software: people make solutions to particular problems they have, and then they put in the work to make that available to the rest of us.
Anyway, one of my gripes over the course of my studies has been that you can’t really run anything in the background in R while you continue to work on another section of code: a lot of the statistical models I’ve been running use MCMCglmm, which is an amazing package but can be pretty time-intensive. Furthermore, my computer at work wasn’t really up to much, so quite often I couldn’t even work on any writing while my analysis was running. In all honesty, I should have looked harder at the options for running R in the cloud (e.g. on Amazon or Domino), but what I really wanted was just something that would give me a little notification when the analysis was done: that way, I could go off and work on something else until the results were ready for me to look at. As it was, there was a lot of clicking back and forth between windows and getting massively frustrated (or going away for half an hour, only to come back and find that I’d specified something incorrectly and it had failed within 10 seconds and needed to be restarted!).
PushoverR takes advantage of the Pushover API, a service that receives messages from an application and sends them out to specified devices. Connelly has created this package so that Pushover receives a message from R, and communicates this to your iPhone or Android device. It’s really easy to set up, so here’s how to get started:
First, you need to sign up for a Pushover account. Head on over to Pushover, and register your email address. You’ll then get your user key (important information number 1!). After confirmation, you need to register your application – in this case, R! Here’s how I set up mine:
Once this is done, you’ll get an application token (important information number 2!). Now, you need to install ‘Pushover’ on your client device – this costs about £3, and is available from the App Store or Google Play:
Now, we’re ready to go! Fire up your favourite R development environment (which should be RStudio, unless you’re a MORON), and get the relevant library installed and called up:
Now, it’s a good idea to set up variables to hold your user ID and application token so that you don’t have to keep pasting them for any message you want to send. To protect myself from anyone here sending me messages to my phone, I’ve replaced the actual tokens with some nonsense:
############## My settings ###########
userID <-"u5schwarzenegger"
appToken <-"a1dontdisturbmyfriendhesdeadtired"
Anyway, we’re now ready to run some analysis – given that my time-consuming code tends to run via MCMCglmm, I’m just going to copy one of the examples given in Jarrod Hadfield’s excellent documentation. Then I’m going to send a simple message to myself that tells me when the analysis is complete; you could, of course, send some of the results in the message, but for now let’s stick to a very basic notification:
We can set this to execute, and – while it runs – go and do something better, like watch youtube vidoes of Arnie violencing up some guys and then zinging their corpses. BLAM
Anyway. What do you know – my phone popped up with a little message!
SWEET.
This is the very basic functionality of PushoverR – check out Brian’s ‘Readme‘ file for some more details of little extra bits. All of the info here was basically taken from that document, but I’ve added some pictures to make it more fun.
Tom Houslay 