Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

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

st: RE: U - Statistics Kernel of order 3

From   Nick Cox <>
To   "''" <>
Subject   st: RE: U - Statistics Kernel of order 3
Date   Thu, 17 Nov 2011 13:59:30 +0000

I count -somersd- (SSC, Roger Newson's site) and -kernreg2- (SSC) as two commands. 

I want to remind David that posters are asked to explain where user-written commands they refer to come from. 

I just want to comment on -kernreg2- (SSC). -kernreg2- was posted in 1999 on SSC in some exasperation because users of -kernreg- (STB) were being repeatedly bitten by bugs that its authors were for some still unexplained reason unwilling to fix.  But that's all history. Any flavour of -kernreg- I take to be superseded long since by the official command -lpoly-. 

So, nobody should be using -kernreg2- unless they are stuck on some Stata about 6 or 7. 

So, an answer to the question is that if -kernreg2- is based on U-statistics of this kind, then presumably -lpoly- is too. I am not fluent in U-statistics concepts or terminology. 


David Ashcraft

I am trying to implement U-statistics with kernel of order 3 in stata for some robust check. There are three different commands that uses the kernel: somersd,and KERNREG2. however, I am not sure what kernel order both packages are using. Please rewcommend, if there is any other command that implements U-Statistic with kerenl order of 3.

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index