# Pierce's criterion

2 messages
Open this post in threaded view
|
Report Content as Inappropriate

## Pierce's criterion

 Hello all, I would like to rigorously test whether observations in my dataset are outliers.  I guess all the main tests in R (Grubbs) impose the assumption of normality.  My data is surely not normal, so I would like to use something else.  As far as I can tell from wikipedia, Peirce's criterion is just that. The data I am interested in testing is: 1) Continuous on the unit interval 2) Discrete 3) Ordinal on 0 6.  If you need more specifics, (1) refers to the gini index of inequality, (2) refers to measures for the number of assasinations, strikes, etc in a country, (3) refers to ranking data of how politically free a country is. Does R do this test? Thanks a lot, and PS I unlike many economists prefer R over Stata R >>>>> Stata! Sincerely, Ryan Murphy -- Ryan Murphy 2012 B.A. Economics and Mathematics 339-223-4181         [[alternative HTML version deleted]] ______________________________________________ [hidden email] mailing list https://stat.ethz.ch/mailman/listinfo/r-helpPLEASE do read the posting guide http://www.R-project.org/posting-guide.htmland provide commented, minimal, self-contained, reproducible code.
Open this post in threaded view
|
Report Content as Inappropriate

## Re: Pierce's criterion

 Determining what is an outlier is complicated regardless of the tools used (this is a philosophical issue rather than an R issue).  You need to make some assumptions and definitions based on the science that produces the data rather than the data itself before even approaching the question of outliers.  What is an outlier for a normal distribution may be reasonable from a gamma distribution and completely expected from a cauchy distribution. See the 'outliers' dataset in the TeachingDemos package, and more importantly the examples in the help page for it, for a demonstration of the perils of automatic outlier deletion. On Wed, Apr 18, 2012 at 4:11 PM, Ryan Murphy <[hidden email]> wrote: > Hello all, > > I would like to rigorously test whether observations in my dataset are > outliers.  I guess all the main tests in R (Grubbs) impose the assumption > of normality.  My data is surely not normal, so I would like to use > something else.  As far as I can tell from wikipedia, Peirce's criterion is > just that. > > The data I am interested in testing is: 1) Continuous on the unit interval > 2) Discrete 3) Ordinal on 0 6.  If you need more specifics, (1) refers to > the gini index of inequality, (2) refers to measures for the number of > assasinations, strikes, etc in a country, (3) refers to ranking data of how > politically free a country is. > > Does R do this test? > > Thanks a lot, and PS I unlike many economists prefer R over Stata R >>>>> > Stata! > > Sincerely, > Ryan Murphy > > -- > Ryan Murphy > 2012 > B.A. Economics and Mathematics > 339-223-4181 > >        [[alternative HTML version deleted]] > > ______________________________________________ > [hidden email] mailing list > https://stat.ethz.ch/mailman/listinfo/r-help> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html> and provide commented, minimal, self-contained, reproducible code. -- Gregory (Greg) L. Snow Ph.D. [hidden email] ______________________________________________ [hidden email] mailing list https://stat.ethz.ch/mailman/listinfo/r-helpPLEASE do read the posting guide http://www.R-project.org/posting-guide.htmland provide commented, minimal, self-contained, reproducible code.