Discussion created by angelag69 on Jan 3, 2014
Hi fellow forum members,

I have used Exploratory Regression to investigate which factors influence the distances at which elephants enter riverine areas in Kruger National Park (South Africa) in relation to the presence of surface water pools. I am more interested in weighing up the relative importance of all the variables than finding one parsimonious model. This is particularly important in my case since many of my variables are colinear and so cannot be included in a single model. Specifically I am testing hypotheses about rainfall, season, gender and whether the waterhole in question is natural or artificial.

What I would like to know is whether I can report the SUMMARY OF VARIABLE SIGNIFICANCE in order to demonstrate the relative importance of all the variables I've tested, rather than to work towards finding a fitting model (given that the VIF is going to be a problem in this case). From all the online tutorials and material I've read from ESRI (which are GREAT, by the way!), it seems I can do this, but since this is for my doctoral dissertation, I'd feel more comfortable with confirmation of this interpretation of mine. My main confusion is whether I can trust the summary if I know that in some cases the 6 checks have not been satisfied.

I'd be very grateful for any feedback/advice.

Kind regards