It looks appropriate for looking at influences on behaviour but we're having trouble finding published studies that use the method. The applied technical pubs on it are less than clear to me in terms of limitations, etc.
some paper citing a paper on beta regression
http://scholar.google.co.in/scholar?hl=en&lr=&cites=17172227630527791062&um=1&ie=UTF-8&sa=X&ei=-dQMUdqkDoqB0AGIw4CIBQ&ved=0CDQQzgIwAA
Thank Saurabh but the ref above seems slightly different since it is dissimilarity to analyze beta diversity. I'm looking at the beta regression that is used when the dependent variable is a proportion (0
You probably want to look for a Logistic Regression. Because of the hard cutoffs, you expect the data to be in the form of a logistic curve.
I know it is a sin, but this is actually a good summary. http://en.wikipedia.org/wiki/Logistic_regression
Some statistical packages like STATISTICA have it build in in to their options.
A place to start with beta-binomial models in R.
http://rss.acs.unt.edu/Rdoc/library/VGAM/html/betabinomial.html
Hello,
We recently used beta regression to model the duration of occupancy (by birds) during a fixed observation period. Please see Lasky & Keitt 2012 Biotropica.
Article The Effect of Spatial Structure of Pasture Tree Cover on Avi...
Ooops, sorry for delay (I've been in the field). Thanks for refs they've been a good start.
Hello Debra,
It is a mystery to my why there are no refs on use of beta distribution, it is clearly applicable for data bounded at both zero and one. CRC publishes many stats books and it turns out they published a book on the beta distribution.
Handbook of Beta Distribution and Its Applications
Arjun K. Gupta, Saralees Nadarajah. 2004 by CRC Press
The price is exorbitant, the Amazon price even higher. Check out the Amazon reviews.
Have a look at:
Generalized Linear Models for Categorical and Continuous Limited Dependent Variables
Michael Smithson, Edgar C. Merkle September 05, 2013
This book covers doubly bounded variables, e.g. bounded at zero and one.
David, thanks much. Doing behavior profiles right can be a nightmare! Will see if my library will spring for it if reviews good. I am in field now so a bit of delay.
Debra,
All I know about Gupta and Nadarajah are the two online reviews and the online Table of Contents. I recommend reading both reviews and the TOC for this book.
Have a look at the Table of contents for Smithson and Merkle. It is way more promising than the TOC for Gupta and Nadarajah. At a way better price.
You can judge for yourself which book is more useful for you, by comparing the TOCs.
~David S
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology