Hi Statalisters,
I am looking for assistance with an analysis project where I would like to multiply impute several nested variables, using the multiply imputed data for the 'higher' variable in my imputation models for the lower variable, in combination with setting some conditional values.
I envision this as employing the following steps:
*1. multiply impute var1
*2. set var2 = 0 for those whose imputed var1==0
*3. multiply impute var2
*4. set var3 = 0 for those whose imputed var2==0
*5. multiply impute var3
*6. combine all multiply imputed datasets for full analysis
All vars are binary.
I am looking for guidance on how to use multiply imputed data to inform the subsequent imputations, how to implement the conditioning based on the mi results and rules, and and how to combine all together for our final analysis.
Thanks in advance for your advice, and where additional clarification is necessary.
~Alison
Related Posts with multi-step data imputation and merging datasets
Help needed: Using lagged explanatory variables as instruments with xtabond2Dear Statalist, I am replicating the paper "Country or leader? Political change and UNGA voting" by…
DiD with (high level) time fixed effects: how to interpret Post coefficient?Hey everyone, Thank you for reading my question. It is related to a "commonly-seen" difference-in-d…
The commands that work in Stata-16 and not in 17.Hello everyone, I used some commands in my do-file in Stata-16 and now they do not work in Stata-17.…
Tabstat tables side by side using esttab in a loopHi everyone, I am trying to get summary statistics of my variables using tabstat and would like to …
Endogenous Switching RegressionHello All, I'm using endogenous switching regression model (cf. Lokshin, Michael and Zurab Sajaia. …
Subscribe to:
Post Comments (Atom)
0 Response to multi-step data imputation and merging datasets
Post a Comment