Re: st: Normally distributed error term & testing normality of residuals

[Nick Cox <njcoxstata@gmail.com>]

See Jay's reply for some good pointers.

Nick

On Sat, Oct 13, 2012 at 2:16 PM, Ebru Ozturk <ebru_0512@hotmail.com> wrote:
> No, I do not refer to that thread.
>
> I need to learn Tobit model so, in order to test whether error term is normally distributed or not what should I do apart from a formal test?
>
> Ebru
>
> ----------------------------------------
>> Date: Sat, 13 Oct 2012 14:05:09 +0100
>> Subject: Re: st: Normally distributed error term & testing normality of residuals
>> From: njcoxstata@gmail.com
>> To: statalist@hsphsun2.harvard.edu
>>
>> The assumption of normality is here associated with a latent variable.
>> If the observed response is bounded, as is the whole point of tobit
>> model, what do you think that implies for the residuals?
>>
>> By the way, if you are referring to the same problem as that discussed
>> in the lengthy thread starting in
>>
>> http://www.stata.com/statalist/archive/2012-10/msg00187.html
>>
>> then my position remains that tobit regression (or more precisely
>> interval regression) is quite inappropriate for that model.
>>
>> Nick
>>
>> On Sat, Oct 13, 2012 at 1:06 PM, Ebru Ozturk <ebru_0512@hotmail.com> wrote:
>> > Thank you Nick.
>> >
>> > I asked it for Tobit model.
>> >
>> > Ebru
>> >
>> > ----------------------------------------
>> >> Date: Sat, 13 Oct 2012 12:42:58 +0100
>> >> Subject: Re: st: Normally distributed error term & testing normality of residuals
>> >> From: njcoxstata@gmail.com
>> >> To: statalist@hsphsun2.harvard.edu
>> >>
>> >> You don't say what kind of model you are fitting.
>> >>
>> >> -qnorm- allows a more incisive graphical assessment of normality than
>> >> -histogram, normal-. It doesn't offer a formal significance test.
>> >> People who are well informed on the matter don't agree: some would
>> >> always seek a formal test, others are happy to think about the results
>> >> of -qnorm- in context.
>> >>
>> >> Even when normality of error terms is an explicit assumption behind a
>> >> model, it is typically also the least important assumption.
>> >>
>> >> Nick
>> >>
>> >> On Sat, Oct 13, 2012 at 12:06 PM, Ebru Ozturk <ebru_0512@hotmail.com> wrote:
>> >>
>> >> > I test for normality of residuals with the command below:
>> >> >
>> >> > predict e, resid
>> >> > histogram e, normal
>> >> >
>> >> > So, does this command also work for understanding of whether error term normally distributed or not?
*
*   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/