xtprobit error

[Enrico Di Bert]

I am trying to work on a xtprobit model with Stata, using the SHARE dataset, which is an unbalanced panel. I am trying to do xtprobit model with smoking as the dependent variable, and quality of life (casp) and education as independent, but I get the error outcome does not vary r(2000). As it is possible to see from the tab, the outcome varies and I have tried aswell to check if there is variation over time for individuals, and I found at least one individual with variation over time. Can you help me?


This are my prompt results:

cd "C:\Users\diber\Desktop\SHARE_Data\sharewX_rel 8-0-0_easySHARE_stata"
C:\Users\diber\Desktop\SHARE_Data\sharewX_rel8-0-0_easySHARE_stata

. use easySHARE_rel8-0-0, clear
(easySHARE release 8.0.0 waves 1 2 3 4 5 6 7 8 doi 6103./SHARE.easy.800)

.
. keep mergeid wave smoking casp eduyears_mod

.
. mvdecode smoking casp eduyears_mod, mv(-1/-20)
smoking: 122038 missing values generated
casp: 66347 missing values generated
eduyears_mod: 37246 missing values generated

.
. bysort mergeid: gen mergeid_n = _n ==1

. replace mergeid_n = sum(mergeid_n)
(412,109 real changes made)

.
. xtset mergeid_n wave
panel variable: mergeid_n (unbalanced)
time variable: wave, 1 to 8, but with gaps
delta: 1 unit

.
. xtprobit smoking casp eduyears_mod
outcome does not vary; remember:
0 = negative outcome,
all other nonmissing values = positive outcome
r(2000);

end of do-file

r(2000);


. tab smoking

Smoke at the present time | Freq. Percent Cum.
--------------------------------------+-----------------------------------
1. Yes | 46,267 15.95 15.95
5. No | 243,805 84.05 100.00
--------------------------------------+-----------------------------------
Total | 290,072 100.00

[Andrew Musau]

The convention in Stata is that \(y=0\) indicates a negative outcome whereas \(y\neq 0\) and nonmissing indicates a positive outcome. The error message tells you exactly this:

outcome does not vary; remember:
0 = negative outcome,
all other nonmissing values = positive outcome
r(2000);

So you need to recode your outcome variable. From

. tab smoking

Smoke at the present time | Freq. Percent Cum.
--------------------------------------+-----------------------------------
1. Yes | 46,267 15.95 15.95
5. No | 243,805 84.05 100.00
--------------------------------------+-----------------------------------
Total | 290,072 100.00

you need

recode smoking (5=0), gen(smoker)
xtprobit smoker ...

See

help recode

[Enrico Di Bert]

Thank you a lot for the help, it was quite a basic mistake and now is working!