I'm using melogit command for a multilevel mixed effect regression. When I run the following command melogit health SEX1 AGE1 Left Right || countryID: I obtain an error message like
"initial values not feasible" r(1400).
Then, I decide to use meqrlogit and it run well.
I'm working in Stata 13.
I applied the following suggestion that I saw in this group: logit health SEX1 AGE1 Left Right
matrix b = e(b)
melogit health SEX1 AGE1 Left Right , from(b, skip) || countryID:

and I obtained the same error message "initial values not feasible"

Can I substitute melogit by meqrlogit ?

Thanks in advance!

ESTIMATION RESULT

melogit health SEX1 AGE1 Left Right || countryID:

Fitting fixed-effects model:

Iteration 0: log likelihood = -16686.915
Iteration 1: log likelihood = -16664.002
Iteration 2: log likelihood = -16663.98
Iteration 3: log likelihood = -16663.98

Refining starting values:

Grid node 0: log likelihood = -16120.406

Fitting full model:

initial values not feasible
r(1400);

end of do-file

r(1400);


meqrlogit health SEX1 AGE1 Left Right || countryID:

Refining starting values:

Iteration 0: log likelihood = -15229.903 (not concave)
Iteration 1: log likelihood = -15188.46
Iteration 2: log likelihood = -15172.542

Performing gradient-based optimization:

Iteration 0: log likelihood = -15172.542
Iteration 1: log likelihood = -15168.007
Iteration 2: log likelihood = -15167.513
Iteration 3: log likelihood = -15167.511
Iteration 4: log likelihood = -15167.511

Mixed-effects logistic regression Number of obs = 32187
Group variable: countryID Number of groups = 19

Obs per group: min = 1002
avg = 1694.1
max = 2852

Integration points = 7 Wald chi2(4) = 41.07
Log likelihood = -15167.511 Prob > chi2 = 0.0000

------------------------------------------------------------------------------
health | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
SEX1 | .0739644 .030306 2.44 0.015 .0145657 .1333632
AGE1 | -.0005852 .0008761 -0.67 0.504 -.0023023 .0011318
Left | .1671068 .040944 4.08 0.000 .0868581 .2473556
Right | -.1103563 .0390669 -2.82 0.005 -.1869259 -.0337866
_cons | 1.46544 .1851746 7.91 0.000 1.102504 1.828375
------------------------------------------------------------------------------

------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
countryID: Identity |
var(_cons) | .6384834 .2098664 .3352461 1.216005
------------------------------------------------------------------------------
LR test vs. logistic regression: chibar2(01) = 2992.94 Prob>=chibar2 = 0.0000