ESA 2008 Annual Meeting Workshop # 5
Quantitative Approaches in Ecological Research and Education Using R
The primary goals of this course are to introduce participants to R - a freely available and very powerful tool for conducting statistical analyses and doing ecological modeling - and to demonstrate the capabilities and tools available in R that are specifically relevant to ecologists. Participants are encouraged to bring their own laptops to step through several exercises on their own.
Purpose:
Links and workshop materials:
What to bring:
-
•a good attitude!
-
•a laptop (don’t forget your power cable!)
If you have never used R, it would be a great idea to spend some time on the R website. You should also see if you can get R installed and running on your laptop. For instructions, see this link. Finally, you might want to download some of the course materials below and peruse them. In particular, you might want to print copies of the slides to bring to the workshop. Copies of the notes will be provided.
Before you come:
1. Use restricted maximum likelihood (REML) for model fits in order to compare variance components of competing mixed effect models. Rationale: Standard maximum likelihood (rather than REML) generates
variance components assuming that the fixed effects parameters are known precisely, rather than estimated. This gives an overly optimistic and narrow estimate of the certainty of the variance
components. (However, if one ever wanted to compare two models with different and nested fixed effect structures using a likelihood ratio test, one must fit the two models with standard maximum likelihood).
2. To create a tilde on a Spanish keyboard, type Alt-126 (holding down the Alt-key and pressing 126). See R-help listserv archive for 12 Aug ('tilde') in the subject line). See also http://www.usefulshortcuts.com/alt-codes/punctuation-alt-codes.php.
Addendum (based on questions/comments that arose during the workshop):
