bbcorrdiff: Bootstrap statistics for the difference of correlation coefficients

Double block bootstrap confidence interval for the difference of two correlation coefficients.

Description

This function computes a double block bootstrap percentile confidence interval and a bootstrap standard error for the difference of two z-transformed Pearson correlation coefficients. The z-statistic is defined as Zd = z1-z2 = atanh(r1)-atanh(r2), where r1 is the Pearson correlation coefficient between columns 1 and 2 of the input matrix and r2 is the correlation coefficient between columns 3 and 4. z1 and z2 denote Fisher's z of r1 and r2 respectively.

References:
-Efron, B. and R.J. Tibshirani (1993): An Introduction to the Bootstrap, Chapman & Hall.
-Bühlmann, P. and M. Mächler (2008): Computational Statistics, Lecture Notes, ETH Zurich.

No Matlab toolboxes are required.

• MATLAB release: MATLAB 7.7 (R2008b)