Uncertainties in the analysis of neutron resonance data

Abstract

Recent analyses of the distribution of reduced neutron widths in the Nuclear Data Ensemble (NDE) and in the Pt isotopes find strong disagreement with predictions of random-matrix theory. These analyses combine the maximum-likelihood method with a cutoff on the reduced neutron widths. We show that the method introduces a systematic error (the "cutoff error"). That error (seemingly taken into account for the Pt data) increases with increasing cutoff and decreasing number of data points. We reanalyse the NDE taking the cutoff error into account. While differing in detail, our results confirm the earlier conclusion that the NDE disagrees significantly from RMT predictions.