Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


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

Re: st: Statistically significant difference in R Squared


From   Aniruddha Rajan <aniruddha.rajan@googlemail.com>
To   <statalist@hsphsun2.harvard.edu>
Subject   Re: st: Statistically significant difference in R Squared
Date   Thu, 23 Aug 2012 20:22:51 +0100

Thanks Dave - I did think about the Chow test but like you say, that tests
for breaks in the conditional mean, variance and/or persistence of a
process. I don't think I'm particularly concerned about whether there's
been a 'break' as such, merely about whether the model has a significantly
higher level of explanatory power over the data during later years.

Regards,
Ani

On 23/08/2012 19:36, "Jacobs, David" <jacobs.184@sociology.osu.edu> wrote:

>Seems like you could use a version of the Chow test, which in time-series
>analyses is based on period interactions.
>
>For example, how about creating a dummy coded "1" for one period and "0"
>for the other and then creating interactions with your explanatory
>variables and this dummy?  Perhaps, I'm wrong about the relevance of this
>test to the question, but if these period interactions are statistically
>significant you then would have grounds for concluding that the
>relationships at issue differ by period.
>
>Anyway if you can succeed in explaining coups with economic variables,
>that would be interesting.
>
>Dave Jacobs
>
>-----Original Message-----
>From: owner-statalist@hsphsun2.harvard.edu
>[mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Aniruddha Rajan
>Sent: Thursday, August 23, 2012 2:18 PM
>To: statalist@hsphsun2.harvard.edu
>Subject: Re: st: Statistically significant difference in R Squared
>
>Thanks for your response Anees. In this case I don't think it is possible
>to look at information criteria or other forms of model comparison as the
>empirical representation of the underlying model does not change - it is
>fixed by theory. Hence the change I am attempting to find a way to
>examine is the model fit in this period versus last period.
>
>Regards,
>Ani 
>
>On 23/08/2012 19:06, "Muhammad Anees" <anees@aneconomist.com> wrote:
>
>>If you have to compare the two models, why not read on on the
>>comparison of nested and non-nested models on the list (by searching
>>the archives) and selecting the best or preferred model using the AIC,
>>BIC or HQIC.
>>
>>Comparisons based on R-squared ( or adjusted R-squared) does not need
>>be tested _Statistically. Simply you can chose a model with higher
>>R-squared ( or adjusted R-squared). Otherwise Nick's suggestion is what
>>seems to be the most appropriate one to me.
>>
>>Best
>>Anees
>>
>>On Thu, Aug 23, 2012 at 11:07 PM, Aniruddha Rajan
>><aniruddha.rajan@googlemail.com> wrote:
>>> Thanks very much for your helpful response Nick. The gist of what I'm
>>>trying to do is to see whether the data I have conforms to a specified
>>>theoretical model to a greater extent during the second period as
>>>compared  to the first. I guess the residual plot should be good
>>>enough. Do any  formal tests for something like this exist?
>>>
>>> On 23/08/2012 17:56, "Nick Cox" <njcoxstata@gmail.com> wrote:
>>>
>>>>I don't think the question makes much sense inferentially for several
>>>>quite different reasons. But looking at the residuals as a function
>>>>of time and seeing whether they vary systematically could be a very
>>>>sensible and helpful check.
>>>>
>>>>Nick
>>>>
>>>>On Thu, Aug 23, 2012 at 5:09 PM, Aniruddha Rajan
>>>><aniruddha.rajan@googlemail.com> wrote:
>>>>
>>>>> I am running a time-series regression model over a 10 year period.
>>>>>I would  be interested in splitting the sample into two five year
>>>>>periods and finding  out whether the model has a statistically
>>>>>significantly higher R^2 during  the second period as compared to
>>>>>the first. Is there a test for this, and if  so is it possible to
>>>>>implement in 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/
>>
>>
>>
>>--
>>
>>Best
>>---------------------------
>>Muhammad Anees
>>Assistant Professor/Programme Coordinator COMSATS Institute of
>>Information Technology Attock 43600, Pakistan
>>http://www.aneconomist.com
>>*
>>*   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   |   Site index