Login  Register

AIC for GLM (or GAM?)

classic Classic list List threaded Threaded
1 message Options Options
Embed post
Permalink
Reply | Threaded
Open this post in threaded view
| More
Print post
Permalink

AIC for GLM (or GAM?)

DP_Sydney
Hi all,

 I understand that using AIC as a model selection tool is not straightforward in SPSS. I wonder if someone can give advice about an analysis I need to do.

     My response variable is a proportion (%time a group of animals spent with another group)
     I have 3 (possibly 5) explanatory variables, all continuous viz year, date, focal group size (possibly also total group size) except one possible binomial variable (breeding vs non-breeding season).

I was considering using general linear modelling for this analysis (using which probability distribution I don't know, nor know how to choose one!). But this assumes linear relationships between independent and dependent variables, which may not be the case. I am now thinking that I should use general additive modelling instead. Comments?

Now to the AIC part. If I use the GLM option then how could I do this in SPSS? If I use the GAM option then what program could I use to perform this analysis (as far as I know SPSS cannot perform GAMs like the one I need)?..............I suspect many of you will suggest R but I have no idea how to use it and am hopeless with scripts.

Advice would be very much appreciated. I'm out of my depth here.

Thanks,
Dean

 I wish to use a model selection procedure to determine the 'best' model, but how would I do this in SPSS? (in