Thanks to Kit Baum, a new version of -margeff- is available from the SSC Archive
(or from my personal website; see the signature below).
To install the update, type
ssc install margeff, replace
Margeff is substantially updated:
- partial effects caused by "small" discrete changes are reported, which are expected to be close or identical
with marginal effects reported by mfx. For continuous variables, the small discrete change is the unit of measurement.
For dummies and counts, the "small" discrete change is 1.
- Some errors reported by users are fixed, for example, ologit and oprobit produced wrong results.
(But the standard errors are different from those reported by mfx... on this issue, I will try to consult
with Stata people)
- not gologit but the new gologit2 is supported.
- a new option constant obtains the baseline predicted value(s) around which partial effects should be
interpreted.
- the code is optimized so that marginal effects at means are calculated faster and in a reliable way
Warning:
My original intention was to stop supporting average marginal effects. Because of the substantial
changes in the structure of the ado file, it might be that average partial effects are wrong, especially when
the dummies option is specified. Therefore, please estimate marginal effects at means.
margeff should be faster than mfx... I will fix bugs soon
Have fun,
Tamas
------------------------------------------------
Tamas Bartus, PhD
Associate Professor, Institute of Sociology and Social Policy
Corvinus University, Budapest
1093 Budapest, Fovam ter 8.
Phone: +36-1-482-5290 Fax: +36-1-482-5226
Homepage: www.uni-corvinus.hu/bartus
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
