# Talk:Winsorized mean

Jump to: navigation, search
WikiProject Statistics (Rated Stub-class, Low-importance)

This article is within the scope of the WikiProject Statistics, a collaborative effort to improve the coverage of statistics on Wikipedia. If you would like to participate, please visit the project page or join the discussion.

Stub  This article has been rated as Stub-Class on the quality scale.
Low  This article has been rated as Low-importance on the importance scale.

My first entry. Hope it's helpful, I am just feeling my way around here. —Preceding unsigned comment added by Plf515 (talkcontribs) 17:13, 24 November 2006

A good start, though IMHO the title should be consistent with that of the article Winsorising. Not that I necessarily prefer 'winsorised mean', I just think it is important to be consistent. Btyner (talk) 21:38, 30 December 2007 (UTC)
I would like to see information in advantages/disadvantages, compared to trimmed-mean —Preceding unsigned comment added by 203.206.162.148 (talk) 05:27, 23 September 2010 (UTC)

## Percentage

I always thought that the example shows a 20% winsorized, as 2 out of 10 are replaced. Nijdam (talk) 18:15, 27 August 2011 (UTC)

## Contradictory

"It involves the calculation of the mean after replacing given parts [...], typically discarding an equal amount of both [...]". Make up your mind: are you replacing (cardinality unchanged) or discarding (cardinality decreased)? Urhixidur (talk) 22:14, 2 March 2012 (UTC)

Some individual values are discarded AND replaced by different values, so no real error, but it could be rephrased. Melcombe (talk) 22:46, 2 March 2012 (UTC)
My point exactly. Thanks for fixing this. Urhixidur (talk) 19:49, 12 March 2012 (UTC)

## Probability distribution

I never heard of the winsorised mean for a probability distribution. Nijdam (talk) 08:43, 25 June 2012 (UTC)

A simple search found this, http://books.google.co.uk/books?id=8f8nBb4__EYC&pg=PA30&dq=%22winsorized+distribution%22&hl=en&sa=X&ei=STnoT-foPITK0QWw9s2uCQ&ved=0CDsQ6AEwAA#v=onepage&q=%22winsorized%20distribution%22&f=false , which is definitely working with a theoretial distribution. The book is Introduction to Robust Estimation and Hypothesis Testing by Rand R. Wilcox. I guess there must be similar stuff elsewhere. Melcombe (talk) 10:17, 25 June 2012 (UTC)

## The example is a bit confusing

The example wasn't really helpful to me. Are we dividing by 10 because the sample count is 10, or because we're taking the 10% Winsorized mean? (which looks rather like 20% (as someone pointed out earlier)).. Ai.unit (talk) 10:00, 25 June 2015 (UTC)