SEM model

[suzy]

Is there an explanation for the slm effect of the model in this picture? For example, do darker areas have more influence on the dependent variable? How to explain the influence of spatial aggregation effect on dependent variables in SEM model?  Thank you.

[Virgilio Gómez-Rubio]

Hi,

Is there an explanation for the slm effect of the model in this picture? For example, do darker areas have more influence on the dependent variable? How to explain the influence of spatial aggregation effect on dependent variables in SEM model?  Thank you.

This is not really an INLA question but I will try to answer it. Those estimates (posterior means?) are part of the linear predictor, hence, they will be influencing the mean (assuming a normal model). So you can expect larger values of the response where you observe darker areas (ceteris paribus). I am not sure what you mean by “the influence of spatial aggregation”. If you mean a change of support (i.e., considering larger areas) then  you will most likely find some differences as you are observing a (spatial) process at a different scale. If you mean interaction between areas, then you need to compute the impacts [1].

Best,

Virgilio

[1] https://www.mdpi.com/2227-7390/9/17/2044