Objects in JWST's mirrors are closer than they appear

Abstract

The James Webb Space Telescope (JWST) has revealed extremely distant galaxies at unprecedentedly early cosmic epochs from its deep imaging using the technique of photometric redshift estimation, with its subsequent spectroscopy confirming their redshifts unambiguously, demonstrating the ability of JWST to probe the earliest galaxies, one of its major scientific goals. However, as larger samples continue to be followed up spectroscopically, it has become apparent that nearly all photometric redshifts at these epochs are biased high with confidence >>99%, for as yet unclear reasons. Here we show that this is the same statistical effect that was predicted in different contexts by Sir Arthur Eddington in 1913, in that there exist more lower redshift galaxies to be scattered upwards than the reverse. The bias depends on the shape of the intrinsic redshift distribution, but as an approximate heuristic, all ultra-high photometric redshift estimates must be corrected downwards by up to one standard deviation.