Confidence interval for a Pearson Correlation Coefficient between two distributions which are not iid

28 Views Asked by At

I have two distribtions $f(x)$ and $f(y)$, and I find the Pearson Correlation Coefficient $\rho$ between them. I want to calculate a confidence interval/uncertainty on the value of $\rho$, and the Wikipedia article for $\rho$ says this can be done using the Fisher transformation, but is only valid if the sample pairs are iid. Is it possible to calculate the confidence interval on $\rho$ if the sample pairs are not iid?