On the Bayesian Analysis of REG Data
York H. Dobyns · 1992
Bayesian analysis may profitably be applied to anomalous data obtained in Random Event Generator and similar human/machine experi- ments, but only by proceeding from sensible prior probability estimates. Unreasonable estimates or strongly conflicting initial hypotheses can project the analysis into contradictory and misleading results. Depending upon the choice of prior and other factors, the results of Bayesian analysis range from confirmation of classical analysis to complete disagreement, and for this reason classical estimates seem more reliable for the interpretation of data of this class.