1996/04/22 by William H. Press
Physics and Astronomy · #astro-ph
published as in "Unsolved Problems in Astrophysics", Proceedings of Conference in Honor of John Bahcall, J.P. Ostriker, ed. (Princeton: Princeton University Press, 1996 [in press]) · 14 pages PostScript includes embedded figures. Paper given at Unsolved Problems in Astrophysics conference, Princeton, April 1995
arxiv created 1996/04/22 · arxiv updated 2009/12/01
To understand their data better, astronomers need to use statistical tools that are more advanced than traditional ``freshman lab'' statistics. As an illustration, the problem of combining apparently incompatible measurements of a quantity is presented from both the traditional, and a more sophisticated Bayesian, perspective. Explicit formulas are given for both treatments. Results are shown for the value of the Hubble Constant, and a 95% confidence interval of 66 < H0 < 82 (km/s/Mpc) is obtained.