Title: Generalized reliability based on distances
Authors: Meng Xu - University of Haifa (Israel)
Philip Reiss - University of Haifa (Israel) [presenting]
Ivor Cribben - Alberta School of Business (Canada)
Abstract: The intraclass correlation coefficient (ICC) is a classical index of measurement reliability. With the increasing prevalence of complex data objects such as curves, images or graphs for which the ICC is not defined, there is a need for new ways to assess reliability. To meet this need, a generalization of the ICC, defined in terms of arbitrary distances among observations, is proposed. The Spearman-Brown formula, which shows how more intensive measurement increases reliability, is extended to encompass the distance-based ICC. A simple bias correction is proposed to improve the coverage of bootstrap confidence intervals for the (classical or distance-based) ICC, and its efficacy is demonstrated via simulations. The proposed methodology is illustrated by analyzing the test-retest reliability of brain connectivity networks derived from a functional magnetic resonance imaging study.