st: Flogit: does it support cluster robust?

[Shonda Naples <liminalzest@gmail.com>]

Hello Statalisters,

I'm running flogit on a cross-sectional dataset.

When I run the following code,

 glm y x1 x2 x3 x4 x5 x6 x7 x8 x9 x10 x3*x9, family (binomial) link
(logit) vce(cluster d) nolog

[d is a place dummy]

Stata outputs the following:



Generalized linear models                    No. of obs      =        52
Optimization     : ML                              Residual df     =        51
                                                              Scale
parameter =         1
Deviance         =  25.43817978                    (1/df) Deviance =  .4987878
Pearson          =  24.12644379                    (1/df) Pearson  =  .4730675

Variance function: V(u) = u*(1-u/1)                [Binomial]
Link function    : g(u) = ln(u/(1-u))                  [Logit]


     AIC             =  .8131172
Log pseudolikelihood = -20.14104679                BIC             = -176.0752



There are 12 explanatory variables. Sample size = 55, but residual df
= 54! When I run flogit with the following code,

glm y x1 x2 x3 x4 x5 x6 x7 x8 x9 x10 x3*x9 d, family (binomial) link
(logit) vce(robust) nolog

I get the expected residual df. I must be missing something. Does
fractional logit not support clustered robust standard errors? Or am I
doing something incorrectly?

Thanks.

Shonda
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/