A Statistical Analysis of the COHERENT Data and Applications to New Physics

TL;DR

This paper critically examines the statistical methods used in analyzing COHERENT neutrino data, emphasizing the importance of proper techniques like Monte Carlo simulations for accurate constraints on new physics.

Contribution

It provides a detailed reanalysis of COHERENT data using appropriate statistical methods and offers tools for testing various new physics models.

Findings

Wilks' theorem does not generally hold for COHERENT data

Monte Carlo methods are necessary for accurate confidence regions

Constraints on neutrino non-standard interactions are refined

Abstract

The observation of coherent elastic neutrino nucleus scattering (CE$\nu$NS) by the COHERENT collaboration in 2017 has opened a new window to both test Standard Model predictions at relatively low energies and probe new physics scenarios. Our investigations show, however, that a careful treatment of the statistical methods used to analyze the data is essential to derive correct constraints and bounds on new physics parameters. In this manuscript we perform a detailed analysis of the publicly available COHERENT CsI data making use of all available background data. We point out that Wilks' theorem is not fulfilled in general and a calculation of the confidence regions via Monte Carlo simulations following a Feldman-Cousins procedure is necessary. As an example for the necessity of this approach to test new physics scenarios we quantify the allowed ranges for several scenarios with neutrino non-standard interactions. Furthermore, we provide accompanying code to enable an easy implementation of other new physics scenarios as well as data files of our results.