Reply To: Report the log likelihood for the negative binomial and cox regression

[Attaullah Shah]

#16587

In the nested regressions, you can add any statistics that is available in the e() macro. This is documented in the help file of asdocx, type help asdocx and then browse to the section 4.10 , you shall find the relevant details and example. Here is one example. Since the log likelihood is reported in the e(ll) macro, we need to add the option stat(ll).

 webuse rod93 , clear
 generate logexp=ln(exposure)
 asdocx nbreg deaths i.cohort , exposure(exp) nest stat(ll)

Table: Regression results

	(1)
	deaths
1b.cohort
2.cohort	-0.268
	(0.724)
3.cohort	-0.457
	(0.724)
Intercept	-2.087***
	(0.512)
/lnalpha	0.594**
	(0.258)
Observations	21
Pseudo R2	0.002
ll	-131.38
Notes: \iStandard errors are in parentheses. *** p<.01, ** p<.05, * p<.1