Error: barstyles[].remake_as_copy: class type not found (when creating box plots)

[Jhanidya Bermeo]

Hi all,

I'm working on creating multiple box plots for a project on student performance on standardized tests. One of the items I'm trying to produce is a matrix of box plots - each box would represent the test scores within a program of a certain academic department. Below is the syntax:

graph hbox toeflcompi if ali_placeinten > 4 & accepted ==1 & planacademicorganization == "Music", by(primaryplancode) over(accepted) over(ali_placeinten)

where variables:
toeflcompi is the test score (continuous)
ali_placeinten is a second test placement (continuous)
accepted is whether or not accepted into department (dichotomous)
planacademicorganization is the department (categorical)
primaryplancode is the program (categorical)

This produces the following error message: barstyles[].remake_as_copy: class type not found

I've been able to successfully create a different matrix with: graph hbox toeflcompi if ali_placeinten > 4 & accepted ==1, by(planacademicorganization) over(accepted) over(ali_placeinten)

Not sure whats going on, though I've narrowed it down to thinking the by clause is the problem. If I change the by to over, I'll get a box plot (not a matrix).

Thoughts?

Thank you!!!

[Jørgen Meisfjord]

Hi. I get the same problem if the by clause creates an empty stratum AND my frequecy weights are missing in the same stratum. Example:

graph hbox MEIS [fweight=pw] , by(SMOKING EDUCATION SEX)

If MEIS and pw are missing for all observations in one of the 24 combinations of SMOKING, EDUCATION and SEX, I get the error message: barstyles[].remake_as_copy: class type not found

If I fix pw so that it doesn't assume missing values, the problem is fixed. My multiple box plots are produced as expected (but the one stratum remains empty, of course).

I see that you don't use weights in your syntax, but maybe this still helps narrow it down.

[Nick Cox]

I'd send a dataset that evokes this problem to StataCorp technical support.