I want to study agreement between 2 quantitative measurements, but I noticed they do not meet the required assumptions of normality and equal variance for ICC.
I read about some non-parametric approaches and one of them is bootstrapping, and I want to know if it would be significant if I calculate the ICC3 of these 2 measurements, and then use bootstrapping to calculate confidence intervals. and if so, how can interpret the results taking into consideration the assumption violation?
Thank you.