[Prev][Next][Index]
statistical help
Hello. I am a graduate student at ASU. I am in the midst of analyzing
data and am not sure how to proceed. Here's my problem in a nutshell. I
censused nine different desert spring areas (riparian habitat) for birds.
I classified the birds as either breeding or non-breeding. From the
censuses, I got an average number of birds per season (spring, summer, and
fall) as well as species richness (how many species throughout each
season, not an average). I collected data for two years. My objective is
to determine which environmental and spatial variables influence both
breeding and non-breeding bird numbers and species richness the most. My
independent variables are a measure of vegetation volume, a measure of
broadleafed species volume, riparian patch size, and distance from the
spring which is being surveyed to the next closest spring area. I
consulted with a statistician who ran a fancy Regression model on SAS.
The output indicated if the independent variables influenced breeding
birds differently than non-breeding birds which was not really my
objective. Would it be valid to treat the breeding birds in a separate
regression from the non-breeding birds?? If I want to determine which
variable influences bird numbers and richness most, would I use a stepwise
multiple regression?? It gets more complicated by the fact that I have
three seasons and two years...can I just get an average of bird numbers
and species numbers across years and not worry about seasons? The
statistician indicated that I would be lossing information if I did
this. It seems to me that I would want four separate regression models
because I have four separate dependent variables, average breeding bird
number, breeding bird richness, average non-breeding bird number, and
nonbreeding bird richness. If you have any advice, I would appreciate
it. Please let me know if anything needs clarification. THANK YOU.
Leigh Rouse
Center for Environmental Studies
Arizona State University
(602) 965-2975