2006A&A...453..797B

Until 6 October 2005 sixteen redshifts had been measured of long gamma-ray bursts discovered by the Swift satellite. Further 45 redshifts have been measured of the long gamma-ray bursts discovered by other satellites. Here we perform five statistical tests comparing the redshift distributions of these two samples assuming as the null hypothesis an identical distribution for the two samples. Three tests (Student's t-test, Mann-Whitney test, Kolmogorov-Smirnov test) reject the null hypothesis at significance levels between 97.19 and 98.55%. Two different comparisons of the medians show extreme (99.78-99.99994)% significance levels of rejection. This means that the redshifts of the Swift sample and the redshifts of the non-Swift sample are distributed differently - in the Swift sample the redshifts are on average larger. This statistical result suggests that the long GRBs should on average be at the higher redshifts of the Swift sample.