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: Matsize Increase Leads to Zero Model Degrees of Freedom


From   Alex MacKay <[email protected]>
To   [email protected]
Subject   Re: st: Matsize Increase Leads to Zero Model Degrees of Freedom
Date   Tue, 3 Sep 2013 10:07:02 -0500

Yes, it is the same model. I've just renamed fips to state_id.

Agreed that it is odd! When stata reports zero model degrees of
freedom, does that indicate that the coefficients were not estimated,
but instead "calibrated" (i.e. the number of linearly independent
observations matches the number coefficients)? I thought that was the
case, but this error seems to indicate otherwise.

Alex

>Hi Alex,
>
>That is very odd. Are you sure that you tried to fit exactly the same
>model? The first model reports "note: 54.fips omitted because of
>collinearity", whereas the second reports "note: 54.state_id omitted
>because of collinearity".
>
>Cheers,
>David
>
>On 1 September 2013 23:27, Alex MacKay <[email protected]> wrote:
>> Dear statalist,
>>
>> I ran into an interesting event in Stata yesterday. I had to increase
>> the matsize for some of my regressions to work, and when I re-ran all
>> of them, some of them that previously seemed fine now ran into an
>> error (with the larger matsize). I've included the log output for both
>> the regression with a matsize of 800 and one with a matsize of 10000.
>>
>> With a matsize of 800, all of the statistics and standard errors are
>> reported. With a matsize of 10000, I get "Warning:  variance matrix is
>> nonsymmetric or highly singular," standard errors are not reported,
>> and the model degrees of freedom are reported as zero. You can see
>> that I am running the exact same regression, as the estimated
>> coefficient is the same and the same fixed effects are excluded. In
>> addition, the root MSE changes as well.
>>
>> Any ideas on why the estimate of the variance matrix would change with
>> a larger matsize when the first was nonbinding (only 599
>> observations)?
>>
>> The regressions are run using -areg- in Stata 12 on a Unix server.
>>
>> Thanks,
>> Alex
>>
>>
>> //Matsize == 800
>>
>> note: 2599.week omitted because of collinearity
>> note: 597.retailer_id omitted because of collinearity
>> note: 866.retailer_id omitted because of collinearity
>> note: 877.retailer_id omitted because of collinearity
>> note: 9101.retailer_id omitted because of collinearity
>> note: 54.fips omitted because of collinearity
>> note: 3997.retailer_id omitted because of collinearity
>> note: 4955.retailer_id omitted because of collinearity
>> note: 7005.retailer_id omitted because of collinearity
>> note: 7599.retailer_id omitted because of collinearity
>>
>> Linear regression, absorbing indicators           Number of obs   =        597
>>                                                   F(  49,     45) =          .
>>                                                   Prob > F        =          .
>>                                                   R-squared       =     0.9256
>>                                                   Adj R-squared   =     0.8695
>>                                                   Root MSE        =     0.3085
>>
>>                       (Std. Err. adjusted for 46 clusters in clusterID)
>> ------------------------------------------------------------------------------
>>              |               Robust
>>     ln_price |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
>> -------------+----------------------------------------------------------------
>>          dummy |  -4.044072   3.152507    -1.28   0.206    -10.39355    2.305404
>>
>>
>>
>> //Matsize = 10000
>>
>> note: 2599.week omitted because of collinearity
>> note: 597.retailer_id omitted because of collinearity
>> note: 866.retailer_id omitted because of collinearity
>> note: 877.retailer_id omitted because of collinearity
>> note: 9101.retailer_id omitted because of collinearity
>> note: 54.state_id omitted because of collinearity
>> Warning:  variance matrix is nonsymmetric or highly singular
>> note: 3997.retailer_id omitted because of collinearity
>> note: 4955.retailer_id omitted because of collinearity
>> note: 7005.retailer_id omitted because of collinearity
>> note: 7599.retailer_id omitted because of collinearity
>>
>> Linear regression, absorbing indicators           Number of obs   =        597
>>                                                   F(   0,     45) =          .
>>                                                   Prob > F        =          .
>>                                                   R-squared       =     0.9256
>>                                                   Adj R-squared   =     0.8695
>>                                                   Root MSE        =     0.2950
>>
>>                       (Std. Err. adjusted for 46 clusters in  clusterID)
>> ------------------------------------------------------------------------------
>>              |               Robust
>>     ln_price |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
>> -------------+----------------------------------------------------------------
>>          dummy |  -4.044072          .        .
>> *
>> *   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/
>*
>*   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/
*
*   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/


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