I want to compare some curves to a 'normality corridor', and I would like to show in what regions (x values) the difference is significant.Take the example below, where I want to compare the dashed lines with the green band:

Curve examples

Each curve represents the median of some values. The curves are not continuous: they only have values at integer x locations, therefore I compare them to the normality corridor at 1, 2, 3, etc.

For the "Width" group I want to show that the difference is significant between all curves and the green band between x=1 and x=3, then only two curves between 4 and 6, then only one curve between 7-8, then two curves etc etc etc...

The problem is that I have 12 groups, each with 3 curves and its own normality corridor, and each of them with 17 levels (integer values of x). That makes for a lot of p-values. Here is what I came up with for the groups in the example, with the number of asterisks representing the p-value, but I think the message can be made clearer:

Table of p-values

TL;DR: How to represent local significant differences between curves and a normality corridor, OR how to represent lots of p-values?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.