Population studies of stellar-mass black hole binaries have become major players in gravitational-wave astronomy. The underlying assumptions are that the targeted source parameters refer to the same quantities for all events in the catalog and are included when modeling selection effects. Both these points have so far been neglected when estimating the orientations of the black hole spins. In particular, the detector-frame gravitational-wave frequency used to define frequency-dependent quantities (e.g., 20 Hz) introduces an inconsistent reference between events at the population level. We solve both issues by modeling binary black hole populations and selection effects at past time infinity, corresponding to the well-defined reference frequency of 0 Hz. We show that, while current gravitational-wave measurement uncertainties obfuscate the influence of reference frequency in population inference, ignoring spins when estimating selection effects leads to an overprediction of spin alignment in the underlying astrophysical distribution of merging black holes.

• Received 12 October 2021
• Accepted 10 January 2022

DOI:https://doi.org/10.1103/PhysRevD.105.024076