Dear all,
I am struggling to conduct a panel model regression for Survey of Consumer Finances (https://www.federalreserve.gov/econres/scfindex.htm) data due to a multiple imputation technique used (each household has 5 imputations). Furthermore, I have constructed the panel using a series of triennial cross-sectional datasets, so plan to use a fixed-effects model to account for heterogeneity.
I have ~20 variables, most of which are dummies and 108,942 observations.
So far, I have been trying to use micombine as this is what is recommended in the dataset's codebook: 'micombine regress 'insert model here', obsid(_mi) impid(_mj) detail'.
The output for this is 'insufficient observations'. But this may be redundant anyway since I don't think you can use micombine with xt commands such as xtreg.
Firstly, does anyone have any advice on how one could tackle multiple imputations without micombine, and secondly, is it possible to apply fixed effects using this method.
I appreciate any help & many thanks in advance!
I am struggling to conduct a panel model regression for Survey of Consumer Finances (https://www.federalreserve.gov/econres/scfindex.htm) data due to a multiple imputation technique used (each household has 5 imputations). Furthermore, I have constructed the panel using a series of triennial cross-sectional datasets, so plan to use a fixed-effects model to account for heterogeneity.
I have ~20 variables, most of which are dummies and 108,942 observations.
So far, I have been trying to use micombine as this is what is recommended in the dataset's codebook: 'micombine regress 'insert model here', obsid(_mi) impid(_mj) detail'.
The output for this is 'insufficient observations'. But this may be redundant anyway since I don't think you can use micombine with xt commands such as xtreg.
Firstly, does anyone have any advice on how one could tackle multiple imputations without micombine, and secondly, is it possible to apply fixed effects using this method.
I appreciate any help & many thanks in advance!
