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: Reported significance levels of parameter estimates wrong?

From   Nick Cox <[email protected]>
To   [email protected]
Subject   Re: st: Reported significance levels of parameter estimates wrong?
Date   Wed, 28 Mar 2012 10:14:56 +0100

I pointed you to a precise place on a web page. It was relevant to
your previous posting, not your main problem.

You said nothing about -outreg2- in your previous posting. -outreg2-
is a user-written command (SSC), as you are asked to explain.

If -outreg2- is not performing as you expect, a minimum expectation
for an answerable question is that you show precisely what you typed
that is relevant to the -outreg2- call and precisely what Stata output

Otherwise your question has the same flavour as "My computer is not
working. Anyone know why?"

The first three paragraphs above all allude to the Statalist FAQ. When
you joined the list, you were asked to read the FAQ before your
posting.  Not studying the advice there given has just made your
question too elusive to be answerable (yet).

By the way, when you do produce precise -outreg2- results, it will be
up to other people to try to answer if they can. I do not use


On Wed, Mar 28, 2012 at 9:47 AM, Benjamin Niug
<[email protected]> wrote:
> Dear Nick,
> I do not understand how the webpage you referred to is related to my question.
> I am using  the -outreg2- command. In its helpfile it is stated that
> "(...), the  default symbol(***, **, *). The significance levels of
> 0.01, 0.05, and 0.10 will be automatically assigned in  that order.
> (...)"
> Best regards,
> Benjamin
> Am 28. März 2012 10:31 schrieb Nick Cox <[email protected]>:
>> If you are expecting that Stata stars results according to attained
>> significance levels, that is not so in this case.
>> Please read
>> Nick
>> On Wed, Mar 28, 2012 at 8:53 AM, Benjamin Niug
>> <[email protected]> wrote:
>>> I am estimating a fixed effects regression calculating clustered
>>> standard errors using the -xtreg var1 var2 var3,fe vce(cluster
>>> variable) - command.
>>> STATA reports the parameter estimates and their standard errors.
>>> However, I am suprised that some of the parameter estimates are not
>>> marked as being signficant - although, when calculating the t-values
>>> manually, they should be.
>>> For example:
>>> 0.0538
>>> [0.0263]
>>> or
>>> -0.0764
>>> [0.0327]
>>> I am really puzzled. Does anybody have an explanation?

*   For searches and help try:

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