Physical Sciences: Statistical analysis of close pairs of QSOs

Abstract

THE observation of close pairs of QSOs with very different redshifts has been suggested to be evidence in support of the noncosmological redshift hypothesis^1,2. The statistical arguments used by Stockton¹ and by Wampler et al.² have recently been attacked by Bahcall and Woltjer³ on the ground that the method used is one of a posteriori statistics. Statistics aside, they have ignored the point that in one case², the evidence is not only based on close proximity of the two QSOs but also on two coincidences in wavelengths in the spectra of the two objects. Although the argument against a posteriori statistics is correct in the absolute sense, this method is widely used in astronomy and has generally been accepted, except in cases in which the hypothesis under consideration has not reached the point of general acceptability. The redshift controversy is a good example of this. Evidence as to the existence of non-cosmological redshifts must, because we have few ways of measuring distance, depend upon finding conflicts between distances derived from the redshifts (assumed cosmological) and distances derived in other ways⁴—either through statistical arguments or by finding luminous connections between pairs of objects. If, as seems to be the case, most workers use the redshift as the prime criterion of distance, and subordinate other methods of measurement to it, no solution to the problem is possible.