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: multicollinearity in regression coeffecient


From   David Hoaglin <dchoaglin@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: multicollinearity in regression coeffecient
Date   Mon, 24 Dec 2012 09:20:10 -0500

David,

It appears from the output that all 85 observations with id==1 have
the same value of x (that may be what you meant when you said, "X is
fixed").  What do you see when you make a scatterplot of y versus x?
(It is a good idea to plot the data before trying to fit a
regression.)

If all the values of x are the same, then x is collinear with the
constant in the regression line.  A regression line summarizes the
relation between change in y and change in x.   If x is constant, the
data provide no information on change in x.

David Hoaglin

On Mon, Dec 24, 2012 at 2:33 AM, David Ashcraft
<ashcraftd@rocketmail.com> wrote:
> Hi Statalist:
>
> I have following regression: y=a0+b1X. The data is time series
> y variable takes on different values, however X is fixed. now when I run my simple regression and I get the following results:
>
> regress y x if id==1
> note: x omitted because of collinearity
>
>       Source |       SS       df       MS              Number of obs =      85
> -------------+------------------------------           F(  0,    84) =    0.00
>        Model |           0     0           .           Prob > F      =       .
>     Residual |  .195687827    84  .002329617           R-squared     =  0.0000
> -------------+------------------------------           Adj R-squared =  0.0000
>        Total |  .195687827    84  .002329617           Root MSE      =  .04827
>
> ------------------------------------------------------------------------------
>            y |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
> -------------+----------------------------------------------------------------
>            x |  (omitted)
>        _cons |   .0017973   .0052352     0.34   0.732    -.0086134    .0122081
> ------------------------------------------------------------------------------
>
> I checked the collinearity via -collin- and results are below:
>
> collin y x if id==1
> (obs=85)
> corr(): matrix has zero or negative values on diagonal
> r(504);
>
> The correlation between y and x is:
>
> corr y x if id==1
> (obs=85)
>
>              |        y        x
> -------------+------------------
>            y |   1.0000
>            x |        .        .
>
> I am confused and need some help here on how to resolve this issue. There are several papers reporting results based on the above regression. Any suggestion?
> Regards
>
> David
*
*   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 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index