Autocovariance without time series data

[Nitsan Machlis]

Hi,
I am working with a large dataset of student grades and am trying to create a regression model that shows the effect of each school on a certain grade.
The regression model should be something like this:

grade = school1 + school2 + school3 + ... + schooln

Each schooln variable is a binary dummy that equals 1 if the observed student belongs to schooln.

I have over 800 schools, so creating such a model manually is not an option.
I understood that the autocovariance function may be related to this issue, but seems to be designed for time series data.
Is there any way to run this regression model automatically without specifiying each school group variable?
Thanks.

[Hemanshu Kumar]

I'm not clear what autocovariance you're thinking of. You say you already have school dummies, and it seems all you are trying to do is include them in your model. If so, all you need is a wildcard to specify the school variables collectively, something like:

regress grade school*

or am I not understanding your problem?

[Nitsan Machlis]

I don't have school dummies, I have a categorical variable with a different number for each school. Is there a way to automatically generate 800+ dummy variables from categorical variables?

[Hemanshu Kumar]

Yes, that is easy. You would need to do something like

regress grade i.school

assuming school is the categorical variable that specifies the school. See also

help fvvarlist