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]

Re: st: mprobit with exclusion restriction

From   saqlain raza <>
To   "" <>
Subject   Re: st: mprobit with exclusion restriction
Date   Wed, 23 Nov 2011 12:16:06 -0800 (PST)

For Mr. Maarten Buis,

I asked a question regarding the problem of -mprobit- and I was not sure about its correctness, that's why I quoted as may be 'silly' question. I thought to get guideline from you because you were the person who responded me about my initial question. In the question, I asked your help and not asked how much you charge for one private question. If I dont want to post any problem on statalist group and think that you can give best advice on that, you should have taken care about that although there is nothing bad in that question. If you could not respond in positive way, you dont have the right to satire.

Saqlain RAZA
PhD Student

----- Original Message -----
From: Maarten Buis <>
Sent: Wednesday, November 23, 2011 11:55 AM
Subject: Re: st: mprobit with exclusion restriction

On Wed, Nov 23, 2011 at 11:25 AM, saqlain raza answered privately:
> I am trying to understand what you said about -mprobit- command and I am
> agree that error correlations are independent. I studied the literature of
> Train (2003) about probit models. He says that
> "With probit models, however, normalization for scale and level does not
> occur automatically. The researcher must normalize the model directly."
> Could you please tell me how can we normalize it? OR is there any command in
> Stata to normalize? OR is it -mprobit- command that automatically normalize
> the scale and level of utility? I am asking you because I do not understand
> these sentences about normalization. And how could we confirm after
> estimation that the errors are really uncorrelated?

I would need to see that quote in context as I can think of multiple
ways in which I could interpret that quote. However, right now I don't
have time to look it up. So, you'll have to hope that somebody else on
Statalist will respond.

> My question may be silly thats why I am trying to write you email and not on
> group. Your suggestions will be appreciated.

Follow up questions are best sent to Statalist and not to individual
members, as is discussed on the Statalist FAQ
<>. Think
about it this way: By sending me a question you ask me to invest my
time to solve your problems. Why would I be more willing to answer
silly questions privately than publicly? The answer is that, unless
you are willing to pay a ridiculously high consulting fee, I am not
willing to answer private questions.

-- Maarten

Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
*   For searches and help try:

*   For searches and help try:

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