Bookmark and Share

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: Wald test in Random Coefficient Model


From   Joerg Luedicke <joerg.luedicke@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Wald test in Random Coefficient Model
Date   Fri, 25 Feb 2011 10:05:49 -0500

On Fri, Feb 25, 2011 at 4:05 AM, Maarten buis <maartenbuis@yahoo.co.uk> wrote:

> The bahavior of this test is not quite what we would expect.
> The problem is that the null hypothesis is on the boundary of
> the parameter space, i.e. the null hypothesis is that the
> country level variance equals 0, and a variance can only be
>>= 0. There is a discussion of that in (Gutierrez et al. 2001).


That's right. I guess the rule of thumb in this case would be to
divide the p-value by 2?

In any case, testing whether some variance is greater than zero or not
is of limited use anyway. I mean, you can always quantify the
variability and go from there. Attaching a p-value to that does not
add a lot of information (if any). Let's take OP's case of 40
countries and some 1000 individuals per country as an example. If you
would find some small variation of coefficient x which is "not
significantly different from zero" would one conclude that this effect
varied slightly across countries but it may be the case that this
merely happened by chance? This does not make a lot of sense,
especially in the case of what is basically a convenience sample.
Those 40 countries basically do not represent anything, they are just
that, those 40 countries.

J.
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


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