This document provides an overview of multiple regression quadratic assignment procedure (MRQAP) for network analysis. It explains that MRQAP is useful when observations are not independent due to network dependencies. The procedure works by permuting the dependent variable many times to build a sampling distribution of coefficients to test significance while preserving network dependencies. Examples show how MRQAP can detect correlations between network variables when they exist versus random permutations when there is no correlation. Functions for performing MRQAP in software packages like UCINET and R are also outlined.