Timeline for Why use bar chart with error whiskers instead of box plot?
Current License: CC BY-SA 4.0
14 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 11, 2022 at 14:44 | vote | accept | Igor F. | ||
| Oct 7, 2022 at 0:32 | comment | added | Glen_b | Many posts on site relate to this question, I suggest trying some searches with suitable keywords | |
| Oct 6, 2022 at 9:00 | history | tweeted | twitter.com/StackStats/status/1577946775226355712 | ||
| Oct 6, 2022 at 6:43 | comment | added | Nick Cox | It's remarkable how many texts (and one has to guess courses too) advise supplementing analysis of variance with a look at box plots. Looking at graphs of distributions beats not looking at them, but as already mentioned in answers at the minimum the box plots should show means too. | |
| Oct 6, 2022 at 6:39 | comment | added | Nick Cox | Three excellent answers to date. I want to flag generally the existence of even more negative views from some statisticians. often to be found by Googling one or more depreciatory terms detonator plot, dynamite plot or plunger plot. For example, simplystatistics.org/posts/2019-02-21-dynamite-plots-must-die and biostat.app.vumc.org/wiki/Main/DynamitePlots I find box plots often oversold, especially for discrete data, or equivalently data with just a few distinct values, where their logic often defeats users, and gives rise to many puzzled threads on this site. | |
| Oct 6, 2022 at 6:21 | answer | added | Stephan Kolassa | timeline score: 5 | |
| Oct 6, 2022 at 6:12 | comment | added | Stephan Kolassa | Incidentally, violin plots (also known as bean plots) convey much more information than boxplots. (If need be, you can overlay a box plot over a violin plot, or even add the individual observations if there are not too many of them - for the small numbers involved here, I would definitely try that and see whether it works well.) | |
| Oct 6, 2022 at 0:51 | history | became hot network question | |||
| Oct 5, 2022 at 22:29 | comment | added | whuber♦ | The very same plot appears at stats.stackexchange.com/questions/591170, which includes more information about its source. | |
| Oct 5, 2022 at 17:34 | comment | added | whuber♦ | Boxplots don't lend themselves to making the comparisons implied by this chart (namely, formal tests of differences of means). | |
| Oct 5, 2022 at 17:28 | history | edited | gung - Reinstate Monica | CC BY-SA 4.0 |
labeled link; edited for English
|
| Oct 5, 2022 at 17:22 | answer | added | gung - Reinstate Monica | timeline score: 9 | |
| Oct 5, 2022 at 17:21 | answer | added | mkt | timeline score: 9 | |
| Oct 5, 2022 at 16:50 | history | asked | Igor F. | CC BY-SA 4.0 |