In this work, we discuss interval estimation for the correlation coefficient, which is an important statistical parameter for measuring the linear relationship between two variables. Four methods for constructing confidence interval are presented: the established method based on the z-transformation, and the latest methods based on generalized pivotal quantities and empirical likelihood. The results of simulation studies of the methods discussed are also presented, comparing their performance in different situations. Finally, the methods are applied to systolic and diastolic blood pressure, where the correlation coefficient is compared between healthy individuals and those diagnosed with hypertension.
|
