Gravitational-wave inference in the catalog era: evolving priors and marginal events

Abstract: As the number of gravitational-wave transient detections grows, the inclusion of marginally significant events in gravitational-wave catalogs will lead to increasing contamination from false positives. In this paper, we address the question of how to carry out population studies in light of the fact that some fraction of marginally significant gravitational-wave events are of terrestrial origin. We show that previously published estimates of $p_\text{astro}$, the probability that an event is of astrophysical origin, imply an effective noise likelihood, which can be used to take into account the uncertain origin of marginal events in population studies. We derive a formalism to carry out population studies with ambiguous gravitational-wave events. We demonstrate this formalism using events from the LIGO/Virgo Gravitational-Wave Transient Catalog 1 (GWTC-1) as well as events from the Venumadhav et al. "IAS catalog." We derive posterior distributions for population parameters and discuss how they change when we take into account $p_\text{astro}$. We provide updated individual-event posterior distributions by including population information.