Normality, sample size, and random variation
These bands show where the ordered values would usually fall if the data really came from a normal distribution with the same sample size.
They are created by repeatedly simulating normal samples, sorting each sample, and recording the middle 95% of the simulated ordered values at each Q-Q plot position.
Rug marks are the small tick marks along the horizontal axis of the histogram. Each mark represents one observed data value.
They help reveal clustering, gaps, repeated values, and possible outliers that may be hidden by the histogram bins.
This controls how many normal samples are simulated to calibrate the Q-Q correlation statistic.
More simulations make the p-value less noisy, but they also require more browser computation.
The simulated p-value is the proportion of simulated normal samples whose Q-Q correlation is as small as, or smaller than, the Q-Q correlation from the current sample.
A small value means that a Q-Q plot this non-linear would be unusual for normal data of the same sample size.