Issues to fix in CEDS Specification
From @kmuhl
- Need to change the names of two long CEDS tables for Stata: k12_student_economically_disadvantaged and k12_school_institution_characteristics. Suggest k12_student_econ_disadvantaged and k12_school_characteristics.
-Some of the CEDS variable names are too long for Stata and get truncated. Just renaming the truncated versions for now.
- Add year to attendance table (It looks like the student attendance records are by year, but there’s no year variable (instead, there are a lot of dupe records)
· Assessment files are missing some key identifiers across the files; they can be merged, but only by assuming each student has exactly one ela and math score per year · Having no missing test scores at all is unrealistic · Also, enrollment patterns are much too pretty; usually students transfer schools and leave with much higher frequency · For the school file, it’s more realistic to have one record per school per year rather than one record per school, though one per school is okay in simplified data · The k12_student_identity table has no information and is not needed · The student attendance file needs a school year variable · We don’t need so many diploma types; 1-3 is more typical, and just one is okay for simplified data · The graduation cohort variable in the enrollment table is deterministic (9th grade cohort plus 3 years) instead of being based on graduation year · There seems to be academic record data only for students who graduate · There’s no IEP data · In general, there are a lot of nuisance variables that don’t vary and lack information; the files would be easier to deal with without them