We measure the void probability function (VPF) of galaxies in the Baryon Oscillation Spectroscopic Survey (BOSS). The VPF provides complementary information to standard two-point statistics in that it is sensitive to galaxy bias in the most extreme underdensities in the cosmic web. Thus the VPF is ideal for testing whether halo occupation of galaxies depends on large-scale density, an effect known as galaxy assembly bias. We find that standard HOD model---one parameterized by halo mass only---fit only to the two-point function, accurately predicts the VPF. Additionally, for HOD models where density dependence is explicitly incorporated, the best-fit models fit to the combination of the correlation function and the VPF have zero density dependence. Thus galaxy assembly bias is not a strong source of systematic uncertaintiy when modeling the clustering of massive galaxies.
Comment: 11 pages, 7 figures