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: Standar probit: how to test normality of the residuals

From   Nick Cox <[email protected]>
To   [email protected]
Subject   Re: st: Standar probit: how to test normality of the residuals
Date   Fri, 23 Mar 2012 12:29:02 +0000

I wonder whether you are confusing the probit model with the linear
probability model.

The probit model does not include a normally-distributed error term,
even in principle.

The simplest definition of residual would be observed - predicted
probability. If observed is 0 or 1 then in practice the predicted
probability is either strictly less than observed or strictly more
than observed. I suppose that collectively the residuals might look
approximately normal, but that's not an assumption of the model.

There are more complicated flavours of residuals, which others will
know more about than I do, but I doubt that they change the main point


On Fri, Mar 23, 2012 at 11:49 AM, Luciana Mèndez <[email protected]> wrote:
> Hy!
> How can I test the normality of the residuals after a standar probit?
> I there any command to run LM test for normality?
> Thank you in advance!
> Luciana
> *
> *   For searches and help try:
> *
> *
> *

*   For searches and help try:

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