'Cases in Use' After Imputation Query
Posted: Mon Jun 27, 2011 2:37 pm
I was wondering if anyone could help to clarify an issue around the number of ‘cases in use’ listed at the bottom of the Equations window after running an analysis using imputed data sets from Realcom.
First, some context. I first ran a growth curve analysis without imputation in MLWIN but listwise deletion led to a large number of missing cases. As a result, using the same model, I then conducted multiple imputation using Realcom (in which all of the variables in my model were used in the imputation). However, after running the analysis with the imputed data sets, at the bottom of the Equations window, it says 6615 of 14340 cases in use. This is a larger number of ‘cases in use’ than I had running the models without imputation but is fewer than I expected (I assumed, perhaps erroneously, that if I imputed in this way that it would say that all 14340 of 14340 cases were in use).
While I followed the instructions from the REALCOM-IMPUTE: multiple imputation using MLwin guide by Harvey Goldstein (August 2010), I am wondering if I have made some error in the imputation or in my analysis because I assumed that it would say that all 14340 of 14340 cases would be in use.
Any comments on whether the incomplete number of 'cases in use' at the bottom of the Equations window is abnormal or not, would be greatly appreciated. If not, any suggestions on how to remedy this would be very helpful.
Many Thanks,
Ryan
First, some context. I first ran a growth curve analysis without imputation in MLWIN but listwise deletion led to a large number of missing cases. As a result, using the same model, I then conducted multiple imputation using Realcom (in which all of the variables in my model were used in the imputation). However, after running the analysis with the imputed data sets, at the bottom of the Equations window, it says 6615 of 14340 cases in use. This is a larger number of ‘cases in use’ than I had running the models without imputation but is fewer than I expected (I assumed, perhaps erroneously, that if I imputed in this way that it would say that all 14340 of 14340 cases were in use).
While I followed the instructions from the REALCOM-IMPUTE: multiple imputation using MLwin guide by Harvey Goldstein (August 2010), I am wondering if I have made some error in the imputation or in my analysis because I assumed that it would say that all 14340 of 14340 cases would be in use.
Any comments on whether the incomplete number of 'cases in use' at the bottom of the Equations window is abnormal or not, would be greatly appreciated. If not, any suggestions on how to remedy this would be very helpful.
Many Thanks,
Ryan