Beginning With Bayesian Stats?

Bayesian statistics have been in use since the 18th century.  However, this statistical methodology has been receiving increased focus in ecology, and I thought it might be helpful to compile a few resources that have been extremely helpful with this branch of stats.

First--the software. In the Lockwood lab at Rutgers University, we utilize OpenBUGS as the software for all of our Bayesian statistics.  OpenBUGS allows the user to customize everything from the data input, to the model's parameters and provides helpful graphical output.  For those who prefer to do their coding in R, the package R2OpenBUGS (developed by Sturtz, Ligges & Gelman, 2005) will be essential.

But really, before the software you must determine how you will be running OpenBUGS on your machine.  If you have a Windows computer, you should be able to download the latest version, and install without a problem.  If you have MacOSX, I am aware of 2 options.  Initially, I ran OpenBUGS on a virtual machine (ie. Parallels or VirtualBox) which essential duplicates a windows desktop on your Mac.  This means you have to have access to download the windows operating system, and also enough computing power so that running two OS at the same time doesn't bog down your machine.  I eschewed this method, opting to follow the extremely helpful tutorial by Dr. David Eagle here.  The tutorial requires a bit of programming through the command line, but ultimately will let you run WinBUGS on your Mac without draining the memory on your computer. 

 

After the installations it's time to Learn OpenBUGS.  In the Lockwood lab, we used Dr. Marc Kery's book Introduction to WinBUGS for Ecologists (2010) as our guide.  The book serves a dual purpose: It introduces to readers to the most frequently used statistical tests through the Bayesian lens, and also provides worked examples as a way to begin using WinBUGS/OpenBUGS in R. With Bayesian stats being applied to myriad ecological studies including SDMs and ENMs, I'm looking forward to continuing analyzing my own data with OpenBUGS.

 

Additional Resources on OpenBUGS and its predecessor WinBUGS can be found here:

www.openbugs.net

[1] Gelfand, A. E. and Smith, A. F. M. (1990) Sampling-based approaches to calculating marginal densities, Journal of the American Statistical Association 85: 398--409

[2] Lunn, D., Spiegelhalter, D., Thomas, A. and Best, N. (2009) The BUGS project: Evolution, critique and future directions (with discussion), Statistics in Medicine 28: 3049--3082.

Rob Crystal-OrnelasComment