STATISTICAL-METHODS FOR INVESTIGATING PERIODICITIES IN DOUBLE-GALAXY REDSHIFTS

We show how to test binary-galaxy redshift data for periodicities against all possible monotone-decreasing distribution functions. The significance of the periodicity in both radio and optical data remains high. We use likelihood methods to compare the chances that the data come from periodic versus Newtonian distributions and find that the optical data are greater-than-or-similar-to 15 times more likely to come from a truly periodic distribution. The same calculation for a new compilation of radio data shows a likelihood ratio of 3. We use Bayesian inference techniques to show that the data suggest a much higher probability for the quantization than has been the case in the past.