Re: st: RE: How to get coefficient and R square from time series regression

[Xixi Lin <winnielxx@gmail.com>]

Hi Joe,

You are right, some ID only has 1 or 2 observations, which render the
error. However, if I delete them, there would be incomplete number
list of IDs, then I can not use forvalues z=1/1894 since there are
some gaps in between them. Do you happen to have a solution to solve
this? Thanks.

Best,
Xixi Lin

On Sat, Sep 21, 2013 at 4:55 PM, Joe Canner <jcanner1@jhmi.edu> wrote:
> Xixi,
>
> One more question: Are there at least two observations for each ID from 1 to 1894?  Linear regression with one independent variable cannot be done with only one observation.  (I'm not sure what effect tsset has in this case; I don't do much with time series.)
>
> In any case, I would recommend, for debugging purpses, removing the -capture- before the -reg-, so that you can see what happens right before you get the "no variables defined error".  My guess is that there is an error in the -reg- statement, which you are not seeing because of -capture- and then when you try to use the _b[] and r[] vectors they do not exist and you get "no variables defined".  (I was able to reproduce this behavior for several kinds of regression errors, including "insufficient" observations.)
>
> Regards,
> Joe
>
> ________________________________________
> From: owner-statalist@hsphsun2.harvard.edu [owner-statalist@hsphsun2.harvard.edu] on behalf of Xixi Lin [winnielxx@gmail.com]
> Sent: Friday, September 20, 2013 3:44 PM
> To: statalist
> Subject: Re: st: RE: How to get coefficient and R square from time series regression
>
> Hi Joe,
>
> When I changed it into forvalues z=1/1894, it works in the beginning,
> and then stops, showing the following errors: no variables defined, do
> you know why?
>
> Thanks,
> Xixi Lin
>
> On Fri, Sep 20, 2013 at 3:01 PM, Joe Canner <jcanner1@jhmi.edu> wrote:
>> forvalues z=1/1894
>>
>> -----Original Message-----
>> From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Xixi Lin
>> Sent: Friday, September 20, 2013 2:59 PM
>> To: statalist
>> Subject: st: How to get coefficient and R square from time series regression
>>
>> Hi,
>>
>> The following is the code I am using to generate coefficient and R square, but it gives errors as:invalid numlist has too many elements
>>
>> gen EQ1= .
>> gen EQ2= .
>> tsset Period
>> foreach z of numlist 1/1894{
>> capture reg NIBE_RS NIBE_RSL if ID==`z'
>> replace EQ1=_b[NIBE_RSL] if ID==`z'
>> replace EQ2=e(r2) if ID==`z'
>> }
>>
>> Does anyone knows how to fix it? Thanks.
>>
>> Best,
>> Xixi Lin
>> *
>> *   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/
>
> *
> *   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/