Dropping NA's

[jleather43]

Hi I'm a beginner with Eviews and I'm using version 7.1. I'm trying to conduct propensity score matching using the following code:

EQUATION SVALEQ.LS _SINGLESEX CONTROLSCOMP
COEF GAMMA =SVALEQ.@COEFS

LOGL PSLGT
PSLGT.APPEND @LOGL LKLHD

PSLGT.APPEND INDEX=0
FOR !J=1 TO !K
PSLGT.APPEND INDEX{!J}=GAMMA(!J)*CONTROLSCOMP(!J)
PSLGT.APPEND INDEX=INDEX+INDEX{!J}
PSLGT.APPEND @TEMP INDEX{!J}
NEXT

PSLGT.APPEND GRES = (_SINGLESEX - @CLOGISTIC(INDEX))
'PSLGT.APPEND LKLHD = _SINGLESEX*LOG(@CLOGISTIC(INDEX))+(1-_SINGLESEX)*LOG(1-@CLOGISTIC(INDEX))

PSLGT.APPEND LOGLK0= -LOG(1+EXP(INDEX))
PSLGT.APPEND LOGLK1= INDEX - LOG(1+EXP(INDEX) )
PSLGT.APPEND LKLHD=@RECODE(_SINGLESEX=0,LOGLK0, LOGLK1)

FOR !T=1 TO !K
PSLGT.APPEND @DERIV GAMMA(!t) GRAD{!t}
PSLGT.APPEND GRAD{!t} = GRES*CONTROLSCOMP(!t)
NEXT


PSLGT.ML(showopts, b, m=1000, c=1e-5)
FREEZE(TABGRAD) PSLGT.CHECKDERIV
FREEZE(TABOUT) PSLGT.OUTPUT
FREEZE(TABCOV) PSLGT.COEFCOV

SERIES PSHAT
SERIES COMSUP
MODEL PSCORE
PSCORE.APPEND PSHAT=@CLOGISTIC(INDEX)

PSCORE.SCENARIO ACTUALS
PSCORE.SOLVE
SMPL @ALL IF _SINGLESEX=1
SCALAR MINPS=@MIN(PSHAT)
SCALAR MAXPS= @MAX(PSHAT)
SMPL @ALL
COMSUP=(PSHAT>=MINPS AND PSHAT<=MAXPS)

STORE INDEX MINPS MAXPS PSCORE PSLGT SVALEQ TABCOV
TABGRAD TABOUT
PAGECREATE(PAGE=REPORT) U 1
FETCH MINPS MAXPS PSCORE PSLGT SVALEQ TABCOV TABGRAD
TABOUT

But I keep receiving an error saying 'missing values in @LOGL series at current coefficients at observations 2 in "DO_PSLGT.ML(showopts, b, m=1000, c=1e-5)". I know that there are lots of missing values in my data set but i'm not really sure how to drop them here as Eviews does not appear to do it automatically in this case.

Many thanks

[EViews Gareth]

Code: Select all

smpl if x<>NA and z<>NA

Will drop all observations for which X or Z have NAs.

[jleather43]

Thanks