Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: ordered logistic regression with endogenous variable
From
"JVerkuilen (Gmail)" <[email protected]>
To
[email protected]
Subject
Re: st: ordered logistic regression with endogenous variable
Date
Thu, 11 Oct 2012 09:37:33 -0400
On Thu, Oct 11, 2012 at 8:30 AM, Justina Fischer <[email protected]> wrote:
> Hi Anat
>
> some practical advise:
>
> Ferrer-i-Carbonell, A. and Frijters, P. (2004), How Important is Methodology for the estimates of the determinants of Happiness?. The Economic Journal, 114: 641–659.
>
> have shown for 10-category life satisfaction data that the bias w.r.t. direction and significance from using OLS in place of an ordered probit estimation is negligeably small.
My dissertation advisor and I were reviewers on that article as I
recall, but we were not so favorable about their claims. I'm not so
sure what you can determine about an estimator from analyzing one
specific dataset as opposed to a careful consideration of the model.
> Hence, I would suggest using ivreg2. Just do not try to discuss the size of the effect, but focus on direction and significance in your interpretation.
I think you can probably get away with -ivreg2- assuming the DV isn't
highly skewed or coarse. However, effect size is still meaningful, or
as meaningful as it ever is for Likert-type scales. Switching to
-oprobit- doesn't all of a sudden make it more legitimate to talk
about the effects. For ordinal data -oprobit- has better statistical
properties and the validity of a linear model is more likely to be
true.
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/