Assess Statement in Proc Genmod.

Sanscrit_2prov — Sun, 14 Jun 2020 20:08:12 GMT

I have run the following model in Proc Genmod:

ods listing gpath="Y:\FES\Graphs";
ods graphics on/reset=index imagename="tangible";
proc genmod data=dist;
class famid;
model tar_depress_w2 = tar_depress_base
prov_prac par_prov_prac
/scale=pearson ;
repeated sub=famid;
assess var=(prov_prac) / resample=10000
seed=603708000 crpanel;
store tangible;
run;
ods graphics off;

For some variables I get a graph from the assess option that looks like the following:

For some variables I get a graph that is like those in the SAS documentation. But for this variable it seems like something is off.

All the variables in my model are continuous. Also I am not sure why the p-value is 1 in the top left hand area of the graph.

Re: Assess Statement in Proc Genmod.

SteveDenham — Mon, 15 Jun 2020 13:34:09 GMT

Well, the graph shows that the estimate is centered on zero, which is probably why the p value =1. I suspect that there is a lack of variability in the response variable after accounting for the other main effects. What sort of results (aside from the assess) do you see with a reduced model, where the only independent variable is prov_prac? The slope estimate for the reduced model might shed some light on what is happening.

SteveDenham

topic Re: Assess Statement in Proc Genmod. in Statistical Procedures

Assess Statement in Proc Genmod.

Re: Assess Statement in Proc Genmod.