Yes Nick, you are right.
Only after her answer I understood.
JR
2008/8/8 Nick Cox <[email protected]>:
> This is going round in circles!
>
> Kieran showed how to get just one residual variable.
>
> Nick
> [email protected]
>
>
> Galina An
>
> Thank you very much, Ricardo, it works well!
> Though it generates several residual variables, which I can collapse
> into 1 afterwards.
>
> Joao Ricardo F. Lima wrote:
>
>> how about:
>>
>> ************************begin example**********************
>> sysuse auto, clear
>> recode rep78 1/2= 3
>> tab rep78 , gen(rep)
>>
>> gen constant = .
>> levelsof rep78, local(levels)
>> foreach levl of local levels {
>> reg mpg rep1 rep2 trunk if rep78 == `levl'
>> replace constant=_b[_cons] if rep78 == `levl'
>> predict x`levl', resid
>> }
>> **************end example************************************
>>
>
>> 2008/8/7 Galina An <[email protected]>:
>>
>>> Great! Thank you very much, Kieran!
>
>>> Kieran McCaul wrote:
>>>
>>>> You need modify the code that Richard suggested earlier.
>>>>
>>>>
>>>> gen r == .
>>>> levelsof region_code, local(levels)
>>>> foreach levl of local levels {
>>>> reg migration t4 t2 t3 if region_code == `levl'
>>>> predict x, resid
>>>> replace r = x if region_code == `levl'
>>>> drop x
>>>> }
>>>>
>>>
>>>> -----Original Message-----
>>>> From: [email protected]
>>>> [mailto:[email protected]] On Behalf Of Galina An
>
>>>> I have another related question. I need to 'predict' residual by
> region.
>>>>
>>>> When I run the following code it only uses the coefficients from the
> last
>>>> regression to predict residuals:
>>>>
>>>> by region: reg migration t2 t3 t4
>>>> predict r, resid
>>>>
>>>> if I try to do the loop, it says that r is already defined:
>>>>
>>>> levelsof region_code, local(levels)
>>>> foreach levl of local levels {
>>>> reg migration t4 t2 t3 if region_code == `levl'
>>>> predict r, resid
>>>> }
>>>>
>
> *
> * 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/
>
--
-------------------------------
Joao Ricardo Lima
Professor
UFPB-CCA-DCFS
+553138923914
-------------------------------
*
* 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/
