Talk:Statistical dispersion

?
In meteorological statistics dispersion is one of the three components of the RMSE:

$$RMSE^2=MBE^2+sdbias^2+dispersion^2$$

see for example Hou et al, 2001:

Hou, Dingchen, Kalnay, Eugenia, Droegemeier, Kelvin K. Objective Verification of the SAMEX ’98 Ensemble Forecasts Monthly Weather Review 2001 129: 73-91

http://ams.allenpress.com/amsonline/?request=get-abstract&doi=10.1175%2F1520-0493(2001)129%3C0073:OVOTSE%3E2.0.CO%3B2

—

I've come across the interpretation of Dispersion as "The second highest value less the second lowest value." Is this a widely used measure? And if so, should it be added to the article? 210.55.130.65 21:01, 10 October 2006 (UTC)


 * I've heard of similar things being used. The idea is simply the range, adjusted for the possibility of two outliers that should be ignored on the grounds that they are atypical. Michael Hardy 22:15, 10 October 2006 (UTC)
 * If there are 2 values, the second highest is the lowest and vice-versa, meaning this is either 0 (iff they are the same) or negative. Alfa-ketosav (talk) 08:11, 1 July 2024 (UTC)

dispersion vs. precision
precision should be mentioned in the article. Jackzhp (talk) 20:43, 20 March 2008 (UTC)

sources of statistical dispersion
"In the physical sciences ... One may assume that the quantity being measured is unchanging and stable" can't be correct. There are too many sources of variability in the physical sciences - ranging from Brownian Motion to fundamental limits to measurability (i.e., Heisenberg's Uncertainty Principle). Mfidelman (talk) 13:27, 2 February 2011 (UTC)