[Limdep Nlogit List] Kernel estimation question

[Jill Caviglia-Harris]

LimDep Users

I have a question about the estimation of the ATT using the kernel
defaults in LimDep.  I have a sample that has a good number of controls 
 Why doesn’t LimDep use all the treated observations when estimating the
ATT using the kernel estimation - shouldn't the kernel utilize all these
obervations?   (In comparison, all are used for the nearest neighbor;
and 12 are dropped when using the radius for this same sample).

-Jill

Jill Caviglia-Harris
Associate Professor
Economics and Finance Department
Salisbury University
Salisbury, MD 21801
(410) 548-5591