Another in the long series of phenomenological papers (e.g. Longo’s spiral galaxy handedness silliness; Land et al.’s CMB anisotropy paper; the Kovetz “giant rings” CMB face-palm; the Webb et al. fine structure dipole) claiming to have discovered remarkable evidence challenging the cherished Copernican principle, or its extension, the cosmological principle. This one is by Singal on astro ph today claiming to have discovered evidence of cosmic-scale anisotropy in the viewing angle distribution of 3C catalogue QSOs across the northern hemisphere.
Singal correctly notes that such an uneven allocation of these 48 QSOs (33 to the 0-12 RA interval, and 15 to the 12-24 RA interval) will occur at only a little more than 1% of the time when drawing from a binomial with p=0.5. BUT, I would caution that the apparent significance of this anomaly must be tempered strongly by two key details of the analytical process. First, the observation that the QSOs were “significantly” unevenly distributed arose after a search across four total object categories (two with slightly fewer objects and one with slightly more), so our (100-1)=98% “significance” in the hypothesis testing framework really becomes ~95% “significance”. Second, the binomial is clearly not the true generative model for QSO distribution, even on large scales, since large-scale clustering means that where we find one bright QSO we are more than usually likely to find a second nearby. If we suppose that one in four of Singal’s QSOs occur in such pairs (which seems like a conservative estimate from inspection of the viewing angle plots; though redshift is omitted) the “significance” drops again to below ~90%. Leaving an interesting observational quirk, perhaps requiring follow up to consider any further sources of extra-binomial variance and/or bias in the completeness of the 3C catalog, but nothing to make us worry about Copernicus.