Dear All, I find that (ssc install) -vce2way- command works in Stata 15.1 but not in Stata 16 in the case of -probit- command. For example, in Stata 15.1,
Code:
. webuse nlswork, clear
(National Longitudinal Survey.  Young Women 14-26 years of age in 1968)

. 
. vce2way probit collgrad ln_wage age grade, cluster(idcode year)

Probit regression                               Number of obs     =     28,508
                                                Wald chi2(3)      =          .
                                                Prob > chi2       =          .
Log pseudolikelihood = -1671.9626               Pseudo R2         =     0.8705

                        (Std. Err. adjusted for clustering on idcode and year)
------------------------------------------------------------------------------
             |               Robust
    collgrad |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     ln_wage |     .12983   .0953884     1.36   0.173    -.0571278    .3167877
         age |  -.0231449   .0041371    -5.59   0.000    -.0312535   -.0150362
       grade |   1.467364   .1345051    10.91   0.000     1.203739    1.730989
       _cons |  -21.51607   2.063507   -10.43   0.000    -25.56047   -17.47168
------------------------------------------------------------------------------
Note: 3885 failures and 0 successes completely determined.
Notes:
    Std. Err. adjusted for 4708 clusters in idcode, AND 15 clusters in year.

    Ignore default Wald chi2(.) and Prob > chi2, or F(.,.) and Prob > F, results above.
      If needed, use command -test- to compute the test statistic and p-value of interest.
In contrast, in Stata 16,
Code:
. webuse nlswork, clear
(National Longitudinal Survey.  Young Women 14-26 years of age in 1968)

. 
. vce2way probit collgrad ln_wage age grade, cluster(idcode year)
4708clustersinidcodeand15clustersinyear invalid name
r(198);
Any suggestions? Thanks.