Assessing Compatibility of Direct Detection Data: Halo-Independent Global Likelihood Analyses

TL;DR

This paper introduces two halo-independent statistical methods to evaluate the compatibility of multiple dark matter detection data sets within a global likelihood framework, accounting for different interaction models.

Contribution

It develops novel techniques for assessing data compatibility in dark matter searches without relying on specific halo models, including a unique best-fit halo function and a plausibility region approach.

Findings

The methods successfully applied to CDMS-II-Si and SuperCDMS data.

The global best-fit halo function is a unique piecewise constant function.

The plausibility region effectively indicates data compatibility levels.

Abstract

We present two different halo-independent methods to assess the compatibility of several direct dark matter detection data sets for a given dark matter model using a global likelihood consisting of at least one extended likelihood and an arbitrary number of Gaussian or Poisson likelihoods. In the first method we find the global best fit halo function (we prove that it is a unique piecewise constant function with a number of down steps smaller than or equal to a maximum number that we compute) and construct a two-sided pointwise confidence band at any desired confidence level, which can then be compared with those derived from the extended likelihood alone to assess the joint compatibility of the data. In the second method we define a "constrained parameter goodness-of-fit" test statistic, whose $p$-value we then use to define a "plausibility region" (e.g. where $p \geq 10\%$). For any halo function not entirely contained within the plausibility region, the level of compatibility of the data is very low (e.g. $p < 10 \%$). We illustrate these methods by applying them to CDMS-II-Si and SuperCDMS data, assuming dark matter particles with elastic spin-independent isospin-conserving interactions or exothermic spin-independent isospin-violating interactions.