Introduction
In 2021, a consensus paper titled "Recommended conventions for reporting results from direct dark matter searches" was published in The European Physical Journal C to standardize how the community reports experimental findings. However, a technical error in the definition of Power-Constrained Limits (PCL) necessitated a 2025 Erratum. This update ensures that the statistical methods used to report dark matter exclusion limits are mathematically sound and consistent with established theory.
The Problem: Statistical Underfluctuations
Direct dark matter detection experiments aim to set "exclusion limits," which define the parameter space (mass and cross-section) where dark matter is unlikely to exist based on observed data. A significant challenge arises when an experiment observes fewer background events than expected due to a random downward fluctuation, known as an "underfluctuation."
If researchers report a limit based solely on this "lucky" data, they may claim to exclude models that the detector lacks the physical sensitivity to probe. Such results are overly optimistic and difficult to reproduce. To mitigate this, the community employs Power-Constrained Limits (PCL), which establish a sensitivity "floor" to prevent reporting limits that exceed the detector's actual capabilities.
Correcting the Definition of PCL
The original 2021 manuscript inadvertently misdefined PCL by using "discovery power" instead of "rejection power." This created a discrepancy between the paper’s formal recommendations and the practical methodologies used by major collaborations such as LUX, PandaX-II, and XENON1T.
The 2025 Erratum replaces the discussion in Section 2.2.1 to correctly define PCL based on rejection power, denoted as $\pi(\mu)$. Rejection power is the probability of rejecting a signal hypothesis ($\mu$) when no true signal is present ($\mu = 0$).
The $-1\sigma$ Convention
The authors recommend a specific threshold for this constraint: a minimal rejection power of $\pi_{crit} = 0.159$. In Gaussian statistics, this value corresponds to the $-1\sigma$ quantile of the background-only distribution. In standard "Brazil band" plots, this threshold aligns with the lower edge of the green band. If a calculated statistical limit falls below the signal strength $\mu_{crit}$ associated with this power, the experiment is advised to report $\mu_{crit}$ instead.
Key Insights and Recommendations
- Alignment with Statistical Theory: The correction realigns community standards with the foundational work of Cowan et al. (2011), ensuring that exclusion limits focus on what a detector cannot see rather than what it might discover.
- Conservative Coverage: The PCL method is intentionally conservative. For very small signals, the method provides a "coverage" of 1, ensuring that true signals are not erroneously excluded in regions of low sensitivity.
- Mandatory Transparency: While PCL is recommended for headline results to avoid over-claiming, the authors emphasize that the unconstrained limit (the raw statistical result) must still be included in data releases. This allows the broader scientific community to see the raw data fluctuations without the applied "floor."
Conclusion
By formalizing the $-1\sigma$ quantile as the standard for power constraints, this erratum provides a consistent framework for the dark matter community. While individual collaborations retain the final say in their reporting, these guidelines protect the field from the pitfalls of statistical underfluctuations and ensure that reported limits accurately reflect experimental sensitivity.
Summary updated