Good Afternoon Statalist,
I am attempting to run a model in Stata 14 using menbreg while using svyset to appropriately adjust for complex survey design, and I'm having no luck getting anything to run once I include a random intercept into my model. My data are panel data with five data points per individual, with varID being the person-level identifier, and my data are in long format. My code looks as follows:
In the code, totalmos is the total number of months in each round of data. For some reason, the first model will run just fine (takes only a few seconds to run), while the second model will not seem to progress at all, although the progress wheel continues to spin.
Does anybody have any suggestions for next steps to try to get this model to run? Has anybody encountered this type of issue before with me commands, where the model seemingly runs forever without any results? I understand that these models are computationally intensive, but it should not take over a week to run, should it? Any help would be much appreciated.
Thanks!
Ryan
I am attempting to run a model in Stata 14 using menbreg while using svyset to appropriately adjust for complex survey design, and I'm having no luck getting anything to run once I include a random intercept into my model. My data are panel data with five data points per individual, with varID being the person-level identifier, and my data are in long format. My code looks as follows:
Code:
use MEPS_long.dta set more off set trace on svyset varpsu, strata(varstr) || varID, weight(longwt) singleunit(centered) *Model 1 svy, subpop(if tabsample==1): menbreg days i.independent AGE income i.OCCCAT i.INDCAT i.REGION year, offset(totalmos) intpoints(3) matrix b0 = e(b) *Model 2 svy, subpop(if tabsample==1): menbreg days i.independent AGE income i.OCCCAT i.INDCAT i.REGION year, offset(totalmos) intpoints(3) from(b0) startvalues(zero) || varID:
Does anybody have any suggestions for next steps to try to get this model to run? Has anybody encountered this type of issue before with me commands, where the model seemingly runs forever without any results? I understand that these models are computationally intensive, but it should not take over a week to run, should it? Any help would be much appreciated.
Thanks!
Ryan