Logistic Regression with Clustering

[Karisma Morton]

Hello users,

I have data for about 15,000 9th graders. I am interested in running logistic models that investigate the impact of a number of independent variables (e.g., ses, grades) on student enrollment in a particular course in the 9th grade (1=enrolled, 0=not enrolled). Naturally, the likelihood of enrollment would be influenced by how these students were clustered both at the middle and high school levels and I would, therefore, like to account for that. Middle schools are not nested within high schools. I would like to run a fixed effects model that would consider these middle and high school clusters through an accurate estimation of the standard errors . However, I am not certain how to go about doing this. Could I accomplish this by simply using the code: logit y x, cluster(middleschool, highschool)? Any suggestions would be appreciated.

Thanks!

[Steve Samuels]

Welcome to Statalist, Karisma!

logit y x, cluster(school_id)

will adequately capture the clustering by school. If you are interested in quantifying variation between- and within- schools, then you need melogit (mixed effect logit). You can add school type as a predictor. However if you also want to assess the relative impact of school characteristics (e.g. size, number of sections), compared to individual characteristics, you need melogit (mixed-effects logit).

[Karisma Morton]

Thank you for your response.

I should clarify that each student in my sample attended 1 middle school and went on to attend 1 high school, and as such, would be associated with two school ids. Therefore, I 'm not certain that clustering on school_id would capture students' classification into a middle school cluster AND a high school cluster. Essentially, I would like to know how certain student-level predictors influence the dependent variable knowing that students belong to two separate clusters that are not nested. Any thoughts on how I could run a fixed effects model that takes these two clusters into account?

[Marcos Almeida]

Karisma, you may also check if a "two-way crossed random effects" melogit wouldn't do the trick in your case.

Best,

Marcos