Skip to main content
9 events
when toggle format what by license comment
Aug 25, 2021 at 12:09 comment added Single Malt What I found most useful is “… 1.5 IQR rule of thumb emerged only after much experimentation. It has mutated in some hands to a hard rule for deleting data points, which was never Tukey's intent”. Not only does this show the origin of the 1.5 magic number (useful to know and bizarrely not mentioned often when it is taught) but it provides practical application of knowing that it does have some empirical rationale but it would be a mistake to apply it too rigidly.
Oct 11, 2018 at 19:06 vote accept Silverfish
Oct 10, 2018 at 18:12 history bounty ended gung - Reinstate Monica
Oct 3, 2018 at 17:53 history edited Nick Cox CC BY-SA 4.0
added 596 characters in body
Oct 3, 2018 at 17:23 history edited Nick Cox CC BY-SA 4.0
added 1485 characters in body
Oct 1, 2018 at 16:50 comment added Nick Cox Thanks! if possession of lines or other markers indicating median and quartiles defines a box, then there were box plots long before Tukey named them, and I am confident that he never claimed otherwise. However, many miniature histories in textbooks and elsewhere seem emphatic on the point; mostly, that's a just a meme repeated without evidence like the story that lemmings jump off cliffs as collective suicide. Many of the alternatives to box plots don't even show a box in any sense, so then the field is wide open to include any graphical representation of univariate distributions.
Oct 1, 2018 at 16:32 comment added Silverfish Really appreciate this answer, thanks Nick - looking forward to the additions about the alternatives and hybrids. I think it is probably fair to say "box plots "and friends" form a "family" of data visualisations, though I don't know what that family should be called
Sep 30, 2018 at 7:53 comment added Nick Cox I have other material on box-percentile plots, mountain plots and other hybrid forms, to be added later.
Sep 30, 2018 at 7:48 history answered Nick Cox CC BY-SA 4.0