Re: [Gretl-users] missing values encountered
On Thu, 19 Jan 2017, John Paravantis wrote:
Trying to run a Poisson regression of count data vs a number of
covariates,
I found out that if a covariate has missing values gretl complains that
missing values were encountered.
Is it possible to instruct gretl to just ignore cases with missing values
listwise?
If not, this is super annoying. I may be missing something.
gretl's Poisson regression generally skips observations where one or
more covariates have missing values, so I guess in your case there
must be a stronger degree of "missingness". Example:
<hansl>
open rac3d.gdt --quiet
list X = 0 SEX AGE AGESQ INCOME LEVYPLUS FREEPOOR FREEREPA \
ILLNESS ACTDAYS HSCORE CHCOND1 CHCOND2
AGE[13] = NA
AGE[999] = NA
INCOME[57] = NA
poisson DVISITS X
</hansl>
Allin Cottrell