Dear all,
My question concerns the interpretation of stset-ting my data.
I have a balanced panel of firms (repeated observations in multiple years with time-varying covariates) and want to analyze firm-level exit. My variable "year" has the scale 2000, 2001, ... and "foundation_year" is the year (e.g., 2000) in which a firm is founded. "firm_id" is the panel identifier of a firm.
When I specify
everything looks fine.
However, I believe that I should be using the - origin - or - enter - option following the threat here https://www.statalist.org/forums/for...ysis-e-g-stcox
When I use
the output tells me that I have 20000 observations out of which 4000 "observations end on or before enter()"
I am not sure what this formulation means. 4000 observations is exactly the number of firms I have, so I suspect it is dropping the first observation in each panel (i.e., they are not used in the analysis). But I don´t know why this is / should be the case. Can someone kindly explain what this output means to me / what is going on here?
Thank you very much!
All the best
Leon
P.S.: If someone has an easy explanation for the difference between the -enter- and -origin- options and knows which one to apply here, I would be very grateful too.
My question concerns the interpretation of stset-ting my data.
I have a balanced panel of firms (repeated observations in multiple years with time-varying covariates) and want to analyze firm-level exit. My variable "year" has the scale 2000, 2001, ... and "foundation_year" is the year (e.g., 2000) in which a firm is founded. "firm_id" is the panel identifier of a firm.
When I specify
Code:
stset year, id(firm_id) failure(firm_exit == 1)
However, I believe that I should be using the - origin - or - enter - option following the threat here https://www.statalist.org/forums/for...ysis-e-g-stcox
When I use
Code:
stset year, id(firm_id) failure(firm_exit == 1) origin(foundation_year)
I am not sure what this formulation means. 4000 observations is exactly the number of firms I have, so I suspect it is dropping the first observation in each panel (i.e., they are not used in the analysis). But I don´t know why this is / should be the case. Can someone kindly explain what this output means to me / what is going on here?
Thank you very much!
All the best
Leon
P.S.: If someone has an easy explanation for the difference between the -enter- and -origin- options and knows which one to apply here, I would be very grateful too.
