Realcomimputeload error: `var’ found where numeric variable
Posted: Fri Sep 21, 2012 12:24 am
Dear colleagues,
I successfully completed MI process (m=10) by RealComImpute. When I tried to load the data to Stata by using “realcomimputeload”, I got the error message in Stata: `var’ found where numeric variable expected.
So, I go back to check the data folder: 10 Stata subdata sets have been successfully created. But the type of some variables (i.e., `var’) is character rather than numeric ones. I open the data set (i.e., either imputeiter100 or imp1.dta), and see some “NA” instead of numbers.
How can I fix it?
I am not sure if it could result from the convergence problems because this MI process ran for 8 days, where there were 35 responses (5 continuous variables, and others 30 are binary or ordinary) and 17 variables without missingness. I set up the number of iterations as the default (n=1000).
Any comments are highly appreciated.
I successfully completed MI process (m=10) by RealComImpute. When I tried to load the data to Stata by using “realcomimputeload”, I got the error message in Stata: `var’ found where numeric variable expected.
So, I go back to check the data folder: 10 Stata subdata sets have been successfully created. But the type of some variables (i.e., `var’) is character rather than numeric ones. I open the data set (i.e., either imputeiter100 or imp1.dta), and see some “NA” instead of numbers.
How can I fix it?
I am not sure if it could result from the convergence problems because this MI process ran for 8 days, where there were 35 responses (5 continuous variables, and others 30 are binary or ordinary) and 17 variables without missingness. I set up the number of iterations as the default (n=1000).
Any comments are highly appreciated.