[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: Re: st: Dependent continuous variable with bounded range

From   "Nick Cox" <>
To   <>
Subject   RE: Re: st: Dependent continuous variable with bounded range
Date   Wed, 16 Apr 2008 18:25:33 +0100

I had a visceral reaction to the binomial, as Pavlos' response is 
in principle continuous (although it's an average of a few integers). 

However, it may well be that the discreteness of the binomial is not all
crucial here, 
rather the shape of its variance function. 

People with a closer knowledge of the literature or a deeper theoretical
may wish to comment. 

The binomial is recommended in


Anders Alexandersson

Pavlos C. Symeou <> wrote originally:
>  I have panel data for 100 firms for five years and I want to examine
>  effects of various variables on "reputation". My variable
>  takes  continuous values in the range of 0-10. Namely, it can take
>  values of 1,2,3 but also of 2.5, 9.6, etc. The values that the
>  "reputation" takes in my sample range between 2.6 - 8.3. Can you
>  advise if I can still use panel OLS estimation for panel data or
>  I use a different model? In essence, my main concern is the
>  of the bounded range of my variable.

There were several responses and clarifications. Nick Cox suggested
> gen repute = reputation / 10
> xtgee repute ..., link(logit) family(<continuous>)

Unfortunately this is not allowed, Nick confirmed to Pavlos:
> You are correct in that -xtgee- (Stata 8) does not support -f(gamma)
> link(logit)- or -f(normal) link(logit)-. I was guessing by analogy
> -glm- which does support those combinations. I am away from base at
> present and unable to check Stata 10. (I assume from your reference to
> Stata 8 manuals that you are using Stata 8. If that is true, it is
> prudent to flag the fact in your postings.)
>  I don't know of any reason why those combinations are not supported
> -xtgee- when they are by -glm-. As a programmer, I am sympathetic to
> explanation of the form "Just didn't think of that" or "Not got round
> that yet".
>  The other link functions are in principle unsuitable, as they pay no
> heed to the range restriction, but you could try
>  -link(identity)- and -f(gamma)- or -f(normal)-.

Are we talking about a fractional response on unbalanced panel data?
In that case, how about using using -glm ..., family(binomial)
link(logit) robust- but separately for each panel? That (or with the
probit link) seems to be what Jeffrey Wooldridge suggests at the end
of this paper:
I'm sure he would be happy to clarify this at 2008 SNASUG, if needed :-)

*   For searches and help try:

© Copyright 1996–2015 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index