Skip to contentSkip to filters

We use cookies for analytics.

Asymptotic theory of the quadratic assignment procedure for dyadic data analysis

Lei Shi

Adobe

Abstract

The quadratic assignment procedure (QAP) is widely used to analyse dyadic data in medical and social sciences: to test the association between two dyadic measurements represented by symmetric matrices, it computes a p-value by permuting the units, equivalently permuting the rows and columns of one matrix together, and its multiple-regression extension is popular in psychometrics. Lei Shi fills the gap in statistical theory for QAP by formulating the network models underlying the various procedures and deriving both the asymptotic sampling distributions of canonical test statistics, using U-statistic theory, and the corresponding asymptotic permutation distributions under strong and weak null hypotheses, using the theory of double-indexed permutation statistics. Together these show that properly studentized statistics make QAP robust: finite-sample exact under the strong null hypothesis while preserving the asymptotic type I error rate under the weak null.

Topics