Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Interpretation of Variace Estimates (xtmixed)


From   Christian Weiß <mail@cweiss.org>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Interpretation of Variace Estimates (xtmixed)
Date   Thu, 17 Sep 2009 22:27:28 -0400

Dear Statalist,


I spent the last time "playing" with my random intercept models,
particularly having  a look at the explained variance which I was able
to todo thx to your support. However, I am facing a new issue now and
any suggestion is highly appreciated.

As mentioned in the book:

Rabe-Hesketh, S., & Skrondal, A. (2008). Multilevel and Longitudinal
Modeling Using Stata. College Station, TX: Stata Press.

on page 101: "[...]. Note that the level-2 variance can increase when
adding level-1 covariates, potentially producing a negative R2 for
level 2", and this exactly is happens in my situation.

If I add several level1, level2 and level 3 covariates to my model,
the R2 for level turns negative. Unfortunately, I am not quite sure
was that meas. For what reason ever, the included covariates "inflate"
the variance of level2. Perhaps, the variables chosen are not very
suitable?

Best
Chris




On Wed, Sep 2, 2009 at 11:42 AM, John Antonakis <john.antonakis@unil.ch> wrote:
> Hi:
>
> Take a look at pages 102-103.
>
> Best,
> J.
>
> ____________________________________________________
>
> Prof. John Antonakis
> Associate Dean Faculty of Business and Economics
> University of Lausanne
> Internef #618
> CH-1015 Lausanne-Dorigny
> Switzerland
>
> Tel ++41 (0)21 692-3438
> Fax ++41 (0)21 692-3305
>
> Faculty page:
> http://www.hec.unil.ch/people/jantonakis&cl=en
>
> Personal page:
> http://www.hec.unil.ch/jantonakis
> ____________________________________________________
>
>
>
> On 02.09.2009 17:21, Christian Weiß wrote:
>>
>> Dear John,
>>
>> browsing to my edition of this book I could not find the respective
>> topic (R-square in multilevel models). Could you provide any more
>> detailed reference to the book?
>>
>> Viele Grüße
>> Christian
>>
>>
>>
>>
>> On Tue, Sep 1, 2009 at 6:13 AM, John Antonakis<john.antonakis@unil.ch>
>> wrote:
>>
>>>
>>> Hi:
>>>
>>> There should also be formulas for r-square, at level 1 and 2 are in:
>>>
>>> Rabe-Hesketh, S., & Skrondal, A. (2008). Multilevel and Longitudinal
>>> Modeling Using Stata. College Station, TX: Stata Press.
>>>
>>> Best,
>>> J.
>>>
>>> ____________________________________________________
>>>
>>> Prof. John Antonakis
>>> Associate Dean Faculty of Business and Economics
>>> University of Lausanne
>>> Internef #618
>>> CH-1015 Lausanne-Dorigny
>>> Switzerland
>>>
>>> Tel ++41 (0)21 692-3438
>>> Fax ++41 (0)21 692-3305
>>>
>>> Faculty page:
>>> http://www.hec.unil.ch/people/jantonakis&cl=en
>>>
>>> Personal page:
>>> http://www.hec.unil.ch/jantonakis
>>> ____________________________________________________
>>>
>>>
>>>
>>> On 01.09.2009 11:11, Maarten buis wrote:
>>>
>>>>
>>>> --- On Mon, 31/8/09, Christian Weiß wrote:
>>>>
>>>>>
>>>>> Please assume xtmixed ,variance yields the following
>>>>> result.
>>>>>
>>>>>
>>>>>
>>>>> ---------------------------------------------------------------------------
>>>>> Random-effects Parameters | Estimate Std. Err. [95% Conf.
>>>>> Interval]
>>>>>
>>>>>
>>>>> -----------------------------+---------------------------------------------
>>>>> Variable1: Identity |
>>>>> var(_cons) | 100.000 70.000 54.22297 415.9787
>>>>>
>>>>>
>>>>> -----------------------------+---------------------------------------------
>>>>> Variable2: Identity |
>>>>> var(_cons) | 10.0000 6.000 4.254477 36.79754
>>>>>
>>>>>
>>>>> -----------------------------+---------------------------------------------
>>>>> var(Residual) | 400.000 40.000 397.8473 465.2598
>>>>>
>>>>>
>>>>> ---------------------------------------------------------------------------
>>>>>
>>>>> Is it 'correct' to conclude that this estimated model
>>>>> yields a total variance of 510 (100+10+400) and
>>>>> accordingly variable1 explains about 20% and variable
>>>>> 2% of the total model or is that a total
>>>>> misinterpretation?
>>>>>
>>>>
>>>> I once knew the answer, now I only know the reference:
>>>> chapter 7 of Tom Snijders and Roel Bosker (1999)
>>>> "Multilevel Analysis, An introduction to basic and
>>>> advanced multilevel modeling". Thousand Oaks: Sage.
>>>>
>>>> Hope this is still of some help,
>>>> Maarten
>>>>
>>>> -----------------------------------------
>>>> Maarten L. Buis
>>>> Institut fuer Soziologie
>>>> Universitaet Tuebingen
>>>> Wilhelmstrasse 36
>>>> 72074 Tuebingen
>>>> Germany
>>>>
>>>> http://home.fsw.vu.nl/m.buis/
>>>> -----------------------------------------
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>     >
>>>> *
>>>> *   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/
>>>>
>>>
>>> *
>>> *   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/
>>>
>>>
>>
>> *
>> *   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/
>>
>
> *
> *   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/
>

*
*   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 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index