[R] Choice of notch size in R
p.dalgaard at biostat.ku.dk
Mon Apr 21 08:08:44 CEST 2008
Alex Reynolds wrote:
> Is there a way to modify the choice of notch size  in R's boxplot
> routine from outlining a 5% significance region, to say 1% or lower?
Not easily. If you look inside boxplots.stats you'll find the hardcoded
constant 1.58, and the documentation has the following text:
The notches (if requested) extend to '+/-1.58 IQR/sqrt(n)'. This
seems to be based on same calculations as the formula with 1.57 in
Chambers _et al._ (1983, p. 62), given in McGill _et al._ (1978,
p. 16). They are based on asymptotic normality of the median and
roughly equal sample sizes for the two medians being compared, and
are said to be rather insensitive to the underlying distributions
of the samples. The idea appears to be to give roughly a 95%
confidence interval for the difference in two medians.
Judging from the wording, either the theory is unclear, or the author of
the help page was not up to speed. My bets are on the former, but as
things stand, the best way forward is to reproduce the calculations
leading up to 1.58 and substitute 0.05 by 0.01 in the appropriate place.
>  McGill, Tukey, and Larsen. "Variations of Box Plots", The American
> Statistician, Vol. 32, No. 1, 12-16.
> R-help at r-project.org mailing list
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
O__ ---- Peter Dalgaard Øster Farimagsgade 5, Entr.B
c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K
(*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk) FAX: (+45) 35327907
More information about the R-help