No evidence for anisotropy in galaxy spin directions

Abstract

Modern cosmology rests on the cosmological principle, that on large enough scales the Universe is both homogeneous and isotropic. A corollary is that galaxies’ spin vectors should be isotropically distributed on the sky. This has been challenged by multiple authors for over a decade, with claims to have detected a statistically significant dipole pattern of spins. We collect all publicly available datasets with spin classifications (binary Z-wise/S-wise), and analyse them for large-angle anisotropies (ℓ ≤ 2). We perform each inference in both a Bayesian and frequentist fashion, the former establishing posterior probabilities on the multipole parameters and the latter calculating p-values for rejection of the null hypothesis of isotropy (i.e. no power at ℓ > 0). All analysis indicate consistency with isotropy to within 3σ. We similarly identify no evidence for a ‘hemisphere anisotropy’ that neglects the angular dependence of the dipole. We isolate the differences with contrary claims in the ad hoc or biased statistics that they employ. Our code is publicly available https://github.com/harrydesmond/GalaxySpinAnisotropy.