Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: Re: A fractional logit can't converge. Anyway, backed up results are useful for mfx? how?

From   Nick Cox <[email protected]>
To   "'[email protected]'" <[email protected]>
Subject   st: RE: Re: A fractional logit can't converge. Anyway, backed up results are useful for mfx? how?
Date   Thu, 8 Mar 2012 12:40:36 +0000

I presume "tobit" is here a typo for "logit". Apart from that, this is often unclear and even contradictory, e.g. over whether you used one predictor or several.

I have two suggestions for your -glm- fitting. Plot the data for the model which didn't converge and see whether the model makes sense. If it does, you may need to tweak the maximize_options. I have had similar difficulties some of the time and found that -technique()- can be crucial. The details are on another computer but my hazy recollection is that -technique(bhhh)- worked better. 

However, for consistency you should refit all your models the same way. 

You should not worry about the note you see, but it is a reasonable message given your use of -family(binomial)-. 

In general, models that do not converge are, in my view, not publishable in any sense, however good the partial results look. I'd say, at a minimum, that it is unethical to quote them without flagging explicitly that they did not converge. 

-outreg- is a user-written program from SSC, as you are asked to explain. 

[email protected] 

Roger Cremades Rodeja

  I am using fractional tobit "glm depvar indepvar, link(logit)
   family(binomial)  vce(robust)". I successfully used it in other two models,
  but in this case after some iterations Stata says "backed up" and cannot
  converge, then I create a table with the results (outreg), and still appears
  a resulting table with results that are consistent with my other two models,
  do those results makes statistical sense?

  I want to create then elasticities with "mfx" but I can't because the
  model did not converge, if those results produced with outreg (previous
  line) make a bit of sense, how can I create elasticities with them using

  Also, after typing the command appear "note: perc_mod_com has noninteger
  values", I do not understand why, as I think proportion variables should be
  0 to 1. Following previous emails in Statalist, I tried delete some
  variables and I still couldn't solve it, OLS do not reject any variable, the
  model do not work trying to delete indep variables one by one, and I am
  trying more and more combinations, but it still does not work.

  Please somebody knows if this analysis can be carried on using another
  sofware? I need these results..

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index