Probit Model - error r(198) - Quadrature points

[Maastricht2006]

Hello,

I want to make a probit model with -xtprobit-, that uses takeover likelihood as dep.variables and governance variables and profitability variables as indep. variables, but I always receive the error message 198, "number of quadrature points must be less than or equal to number of observations".
I could not find out the source of the problem with the available help tools. Any ideas?
I think I have sufficient observations, but I am not sure about quadrature points, it is related to numerical integration

Stata:
xtprobit takeovercompleted classifboard limabmeeting limabconsent goldenparac
> hute poisonpill ebit ebitda netincome netcashflow
number of quadrature points must be less than or equal to number of obs
r(198);


Interestingly, if I use less indep. var. I receivee no error and receive this Stata output:

xtprobit takeovercompleted classifboard limabmeeting ebit ebitda

Fitting comparison model:

Iteration 0: log likelihood = -2497.4917
Iteration 1: log likelihood = -2495.9241
Iteration 2: log likelihood = -2495.9241

Fitting full model:

rho = 0.0 log likelihood = -2495.9241
rho = 0.1 log likelihood = -2466.412
rho = 0.2 log likelihood = -2462.8364
rho = 0.3 log likelihood = -2474.9509

Iteration 0: log likelihood = -2462.8363
Iteration 1: log likelihood = -2461.5604
Iteration 2: log likelihood = -2461.5525
Iteration 3: log likelihood = -2461.5525

Random-effects probit regression Number of obs = 3606
Group variable: gvkey Number of groups = 942

Random effects u_i ~ Gaussian Obs per group: min = 1
avg = 3.8
max = 11

Wald chi2(4) = 1.89
Log likelihood = -2461.5525 Prob > chi2 = 0.7551

------------------------------------------------------------------------------
takeoverco~d | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
classifboard | -.0183734 .0547695 -0.34 0.737 -.1257197 .0889728
limabmeeting | .0621905 .0537666 1.16 0.247 -.0431901 .1675711
ebit | -.00002 .0000342 -0.58 0.559 -.0000869 .000047
ebitda | .0000131 .0000287 0.46 0.648 -.0000431 .0000693
_cons | .0109306 .0453327 0.24 0.809 -.0779198 .099781
-------------+----------------------------------------------------------------
/lnsig2u | -1.583594 .1722098 -1.921119 -1.246069
-------------+----------------------------------------------------------------
sigma_u | .4530299 .0390081 .3826787 .5363145
rho | .1702871 .0243314 .1277368 .2233813
------------------------------------------------------------------------------
Likelihood-ratio test of rho=0: chibar2(01) = 68.74 Prob >= chibar2 = 0.000

[bukharin]

This is most likely caused by missing data in your predictor variables leading to cases being dropped. Try:
count if !missing(gvkey, takeovercompleted, classifboard, limabmeeting, limabconsent, goldenparachute, poisonpill, ebit, ebitda, netincome, netcashflow)

[Maastricht2006]

/lnsig2u | -1.583594 .1722098 -1.921119 -1.246069


does somebody know what "lnsig2u" means exactly? It refers to the stata output above