When groups contain groups: measuring hidden structure in networks

Most network science treats connections as simple pairs, but real interactions often involve groups within groups — a lab within a department, a clique within a community. This review unifies several independent measures of this 'nestedness' under a single mathematical framework called the encapsulation graph. Across social datasets, nestedness turns out to be common — and strikingly, its absence is just as informative, signaling distinct mesoscale structure in the system.