Simultaneous confidence intervals

Open aphalo opened this issue 2 years ago • 4 comments

A statistic could be rather easily implemented by editing the code of stat_multcomp().

Aug 13 '23 13:08 aphalo

Closed by mistake.

Nov 12 '23 13:11 aphalo

For clarity, would this be a confidence interval for the delta between groups?

Nov 12 '23 20:11 markbneal

I haven't really thought the details. What would be most useful? I have been playing with the idea in my head, as I find the p-value labeling using horizontal segments a bit counterintuitive. Wouldn't it be better to have a vertical bar highlighting the actual difference along the y-axis, and somehow highlight its confidence? I raised this issue mostly to come back at some time to this.

Nov 13 '23 22:11 aphalo

An interesting alternative way of plotting is described and implement in 'emmeans' under the name "Pairwise P-value plots".

Apr 15 '24 21:04 aphalo