Date: Wed, 24 Feb 1999 16:55:39 -0500
Sender: "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From: Justin Quirouette <Justin_Quirouette@PCH.GC.CA>
Subject: Monte Carlo for Logistic Regression?
Content-type: text/plain; charset=us-ascii
Please forgive any innapropriate use/abuse of terms (newbie).
I have run a logistic regression model based on presence/absence data of a
species distribution. What I would like to do is somehow compare my species
model results with a model based on completely random presence and absence data
to see if there is a significant difference. My reasoning is that a model
predicting a species distribution should produce a much more accurate result
than a model using completely random values.
My understanding is that a Monte Carlo simulation could/should be performed
whereby the predictor variables are reassigned new values and the logistic
regression model is re-run several times. Does anyone know of any scripts that
would automate such a task, or any other concerns, considerations, or
appropriate solutions (please remember the newbie factor)? I am running SPSS
Any and all information will be greatly appreciated. Will SUM the good stuff.
Thank you for your time,