Problems with mi extract. All imputed variables have the same value.

[Filippo Mauri]

Dear statalists.
I have succesfully performed a multiple chained imputation on my model ( 300 banks in which I'm analazying variablies like revenue, debt eccto predict the probability in a spike of the price via logistic rgression)
The problem is that when I save mi.dta , open it a try to extract the databases( mi extract n) all databases have the same imputed values. I know it's unconvential to extract and analyse single imputed database,but that's what i need to do.
Thank you!

[Rich Goldstein]

to help with the specific question asked, more is needed (e.g., exactly what command you gave to Stata) - please read the FAQ

but also note that you may be able to do what you want use -mi req-; see

help mi xeq

[Filippo Mauri]

Thank you richard.here is the code

mi register imputed vardepreciation DepreciationPropertyplanteq varDepreciationPropertyplant capitalexpendituretotalassets varLTDebttoTotalEquity FreeCashFlowtoTotalDebt repaymentlongtermdebtlong issuancelongtermdebtlongt varNetChangeLongTermDebt DIVIDENDFREECASHFLOW Tier1CommonEquityRatio RiskWeightedAssets VarRiskWeightedAssets TotLiabCommonEquity interestbearingliabilitiestot TotalRiskBasedCapitalRatio LTInvestmentsReceivables shorttermdebttotaldebt timedeposittotaldeposit NonPerformingLoanstoTotalLo varNonPerformingLoanstoTot prestiticommercialiprestitito varProvisionforLoanLosses varPersonnelExpenses varDeferredIncomeTax BankFeeMargin varBankFeeMargin








mi impute chained (regress) vardepreciation Depreciati~q varDepreci~t capitalexp~s varLTDebttoTotalEquity FreeCashFlowtoTotalDebt repaymentlongtermdebtlong issuancelo~t varNetChangeLongTermDebt DIVIDENDFREECASHFLOW Tier1CommonEquityRatio RiskWeightedAssets VarRiskWeightedAssets TangibleCo~W interestbe~t TotalRiskB~o LTInvestme~s shorttermd~t timedeposi~t NonPerform~o varNonPerf~t prestitico~o varProvisi~s varPersonn~s varDeferre~x BankFeeMar~n varBankFee~n = binariaMEDIA varrevenue returncommeq I LTDebttoTo~y Equityfixe~t varEquityf~t salestotal~s varsalesto~s retutnonas~t opprofitto~s varopprofi~s pretaxinco~s varpretaxi~s profitmargin varprofitm~n salescash salesfixed~t vartotalas~s OPERATINGI~S varOPERATI~T varTotalfu~s cashflowne~e Tier1Capit~o Vartier1ca~o CommonEqui~s VarCommonE~s TotLiabCom~y TangibleCo~o Tangibleeq~y CashandMar~t varCashand~t ShortTermI~s WeightedAv~p varWeighte~a varActualC~e varTotalDe~s , add(5) rseed(2222)


after this, I saved the file with: save mi
I opened the file, extracted one imputation with mi extract n. closed the file and reopened it and extracted another imputation. they are all the same.


More clear?

[Rich Goldstein]

if you really typed "mi extract n" twice, then, (1) I'm not sure what Stata did, but (2) whatever it did was the same each time so it's not surprising that the results are the same

[Filippo Mauri]

nono.
example:
mi save
mi extract 1

AFTER THIS I CLOSE THE SESSION, RE OPEN MI SAVE.DTA, and extract another imputation.( for example mi extract 2, which should give me the m=2 imputed database)
is that clear?
btw If you have any idea how to extract the different database, I'd be really thankfull.
Thank you

[Filippo Mauri]

Dear richard,
I used the code you suggested to summarize my dataset and I can see that the value I have imputed have different means, so it seems i'm unable to extract different database due to an error.
I wait for you kind response