Nick Cox replied:
> The argument for logit is correct in principle, but over
> the range from 0.14 to 0.15 logit of a proportion is as near
> linear as is needed for almost all practical purposes.
> In fact, forget the "almost". This really is a detail
> compared with others.
>
> If you are going to transform, note that Stata has a -logit()-
> function. I prefer to do it by -glm-:
Good catch: forgot to mention that. I note that by fitting
> . glm index year, link(logit)
and getting
> year | -.006451 .0013616 -4.74 0.000 -.0091197 -.0037822
> --------------------------------------------------------------------------
it matches the intepretation I gave after fitting -reg, eform()-, although
it's not an exact match as the two models are not identical:
. reg logindex year, eform(OR)
[...]
year | .9935576 .0013462 -4.77 0.001 .990458 .9966668
----------------------------------------------------------------------------
. display .9935576-1
-.0064424
The advantage of -glm-, of course, is that you can get there in one
without any faffing about with logit transformations, so I second Nick's
recommendation of it. However, I'm pretty sure that -glm- does suffer
(slightly) in comparison to -reg- as far as postestimation commands are
concerned. But, then, so do most other regression routines.
> Yoking Herfindahl and Hirschman is not appropriate here
> as their measures differ. Herfindahl's measure, or its
> complement, is also known as the Gini index (one of several),
> heterozygosity, Simpson's index, etc.
I blame another Cox (1997: 29) for that. No relation, I trust.
CLIVE NICHOLAS |t: 0(044)7903 397793
Politics |e: [email protected]
Newcastle University |http://www.ncl.ac.uk/geps
Reference
Cox GJ (1997) MAKING VOTES COUNT: STRATEGIC COORDINATION IN THE WORLD'S
ELECTORAL SYSTEMS, Cambridge: Cambridge University Press.
*
* 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/
