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 at the end of May, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: Point Estimates with Local Polynomial Regression


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: Point Estimates with Local Polynomial Regression
Date   Tue, 27 Jul 2010 19:41:22 +0100

One safe answer to the second is surely to -drop- every observation not
in the desired group, even if only temporarily. Then there is _no_
possibility of observations not in the desired subset influencing the
smooth. 

Another more general answer is that I don't ever think of point
estimates from a nonparametric smooth as correct in any strong technical
sense. The choice of kind and degree of smoothing is always capricious,
even when it is automated. 

Nick 
n.j.cox@durham.ac.uk 

Rodney Hughes

With complex survey data, "subpop" should be used instead of "if" in
Stata to account for groups appropriately and calculate standard
errors correctly, but I want to run nonparametric regressions using
Stata's "lpoly" command for local polynomial regression, and lpoly is
incompatible with subpop in Stata 10 and 11.

If I use "if" instead of subpop (say, to limit a nonparametric
regression to observations from only one state), will the point
estimates of the smoothed values in the nonparametric regression be
correct? I know the point estimates would be correct and the standard
errors would be incorrect for other regression models, but I wonder
about nonparametric regression because the smoothed values rely on
other nearby values, and "if" would exclude some of these.


*
*   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