Dear statalists，
Sorry to disturb you again!I face some difficulties when I use gllamm command to do regereesion.Details are as following:


gllamm yyzns comecity_hp,i( comecity_code ) base(0) link(mlogit) family(binomial) adapt nolog
insufficient observations
r(2001);


DV:yyzns IV:comecity_hp,approximately 4456 obsevations and 159 groups
When I do null model:gllamm yyzns,i( comecity_code ) base(0) link(mlogit) family(binomial) adapt nolog,
the output turned to be great.


number of level 1 units = 4456
number of level 2 units = 159

Condition Number = 49.373262

gllamm model

log likelihood = -3346.3716

------------------------------------------------------------------------------
yyzns | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
c1 |
_cons | 5.209611 .6673054 7.81 0.000 3.901717 6.517506
-------------+----------------------------------------------------------------
c2 |
_cons | 6.319957 .6668244 9.48 0.000 5.013005 7.626908
-------------+----------------------------------------------------------------
c3 |
_cons | 3.669485 .6699283 5.48 0.000 2.35645 4.98252
------------------------------------------------------------------------------


Variances and covariances of random effects
------------------------------------------------------------------------------


***level 2 (comecity_code)

var(1): 1.936165 (1.3356846)
------------------------------------------------------------------------------

I would be appraciate it if anyone of you could help me,many thanks!