2
$\begingroup$

I have pupils clustered within schools, with a binary outcome (pupil experiences bullying or does not). So far all predictors are pupil-level data, and I run a the same model (same predictors) on each year of data. My teacher was talking about adding a school-level predictor: the size of the school, which will simply be the count of pupils in each school, that is the number of observations in each school. Is there any problem with doing this ? I tried to look for some references on this but I didn't find anything useful so far, and looking in some textbooks it doesn't seem like a normal thing to do ?

Edit: I am interested in how the residual variance at the school level changes over time.

$\endgroup$
2
  • $\begingroup$ You can certainly have group (school) level predictors in the model - is that what you're asking? $\endgroup$
    – Macro
    Jun 30 '12 at 6:01
  • $\begingroup$ @Macro I know I can have school-level predictors in the model. My understanding of such a predictor is that it is one that does not change at the pupil level, only at the school level (so it's constant across all pupils in each school). My question is whether having the count of observations at each school in the data (which is a school level predictor by the definition I just mentioned), makes sense statistically. For some reason it just seems a bit weird to me. $\endgroup$
    – Joe King
    Jun 30 '12 at 9:35
0
$\begingroup$

Why not just try it? I can see no problems with school size as a predictor, and in studying bullying it might just be that small and large schools are different ...

You say yourself

My understanding of such a predictor is that it is one that does not change at the pupil level, only at the school level (so it's constant across all pupils in each school)

and school size is clearly such a predictor---it is the same for all the pupils.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.