I am working with a database hosted at a secure platform at Statistics Denmark, so there is limits to how much detail I am allowed to share.
The database consist of two cohorts: 1 - all psychiatrically treated eating disorder cases over the years 1970-2014 N: 22,374 , and 2- a 1:4 matched non-exposed cohort (N: 83,060). In all 37 variables and 1 670 017 rows of observations.
I am looking into educational attainment - highest educational level completed. Research question is do the persons exposed to eating disorders experience a delay in educational attainment, or are they distributed differently than the controls ten years after first treatment for eating disorder. Dependent variable is ISCED , a 6-level ordered scale.
I have organized the data for ordered logistic regression - ologit, and I am able to run the initial analyses for various models, but when I try to run ologitgof Stata seems to be overloaded - it runs for ever.
Is there some limits I am not aware of?
Statistics Denmark use version 17.0 MP multi-core etc so there should not be capacity problems. Contact with the support staff revealed that ologitgof is correctly installed.
At home I use stata 17.0 SE, and I can without problems repeat the examples from The Stata Journal (2017), 17, Number 3, pp. 668-686 - by Morten W Fagerland and David W Hosmer
I will b e very grateful for any help.
Søren Nielsen
The database consist of two cohorts: 1 - all psychiatrically treated eating disorder cases over the years 1970-2014 N: 22,374 , and 2- a 1:4 matched non-exposed cohort (N: 83,060). In all 37 variables and 1 670 017 rows of observations.
I am looking into educational attainment - highest educational level completed. Research question is do the persons exposed to eating disorders experience a delay in educational attainment, or are they distributed differently than the controls ten years after first treatment for eating disorder. Dependent variable is ISCED , a 6-level ordered scale.
I have organized the data for ordered logistic regression - ologit, and I am able to run the initial analyses for various models, but when I try to run ologitgof Stata seems to be overloaded - it runs for ever.
Is there some limits I am not aware of?
Statistics Denmark use version 17.0 MP multi-core etc so there should not be capacity problems. Contact with the support staff revealed that ologitgof is correctly installed.
At home I use stata 17.0 SE, and I can without problems repeat the examples from The Stata Journal (2017), 17, Number 3, pp. 668-686 - by Morten W Fagerland and David W Hosmer
I will b e very grateful for any help.
Søren Nielsen
