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: Missing F value in regression output.


From   Steven Samuels <sjsamuels@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Missing F value in regression output.
Date   Thu, 30 Jun 2011 09:11:49 -0400

Sumedha:

As to what your problems implies for your results: if we can't see the results we can't guess. We need to see everything that Stata types, including the regression output (this is an FAQ). It also depends on the purposes of your analysis.

Why is the F statistic missing?  I'm guessing that the variance-covariance matrix is not of full rank.  The clue is that the F numerator d.f. is less than number of regressors. One common reason in the survey setting will not apply here, namely that the number of coefficients is greater than the design degrees of freedom. A more likely reason would be be near-collinearity for the predictors (not perfect collinearity, or Stata would simply drop some of them). You can investigate these things directly. 

*************
mata
v = st_matrix("e(V)")
rank(v)   //rank of the variance-covariance matrix
end
estat vif
*************

By the way, the sample size in your subpopulation is the same as that in the entire sample; is this what you intended? 

Steve

Thank you for your prompt reply and apologies for omitting the syntax.
I had typed:

eststo BMImsw:  svy, subpop(pop2w): reg logYw4 /*incW4*/ BMI currwkW4 MDcurrwkW4 tenW4 MDtenW4  malew1 ///
blackhw1 hisphw1 raceother ageW4 agesqW4 totchild MDtotchild peduyw1 MDpeduyw1 eduyw4 /*lowbrthwt*/ exercisew4 ///
hlthinW4 vocabStdW1 region1 region2 region3 depressw4 marriedw4  

Many thanks,
Sumedha.


________________________________________
From: owner-statalist@hsphsun2.harvard.edu [owner-statalist@hsphsun2.harvard.edu] on behalf of Nick Cox [njcoxstata@gmail.com]
Sent: Wednesday, June 29, 2011 2:58 PM
To: statalist@hsphsun2.harvard.edu
Subject: Re: st: Missing F value in regression output.

You did not specify the syntax you typed, presumably some variant of
-svy:-. -svy- experts are no fonder of guessing games than anyone
else!

Nick

On Wed, Jun 29, 2011 at 7:32 PM, Gupta, Sumedha <sugupta@iupui.edu> wrote:

> I will really appreciate your help in the following:
> 
> For a regression I am running the output begins as follows:
> 
> Number of strata   =         4                  Number of obs      =      3395
> Number of PSUs     =       132                Population size    = 7364711.9
>                                                       Subpop. no. of obs =      3395
>                                               Subpop. size       = 7364711.9
>                                               Design df          =       128
>                                               F(  20,    109)    =         .
>                                               Prob > F           =         .
>                                               R-squared          =    0.1775
> 
> 
> I am not sure why the F(20,109) stat is missing? Also what does this imply for my results?

*
*   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