Abstract
The field of dark matter detection is a highly visible and highly competitive one. In this paper, we propose recommendations for presenting dark matter direct detection results particularly suited for weak-scale dark matter searches, although we believe the spirit of the recommendations can apply more broadly to searches for other dark matter candidates, such as very light dark matter or axions. To translate experimental data into a final published result, direct detection collaborations must make a series of choices in their analysis, ranging from how to model astrophysical parameters to how to make statistical inferences based on observed data. While many collaborations follow a standard set of recommendations in some areas, in other areas, particularly in statistical inference, they have taken different approaches. We set out a number of recommendations on how to apply the now commonly used Profile Likelihood Ratio method to direct detection data. In addition, updated recommendations for the Standard Halo Model astrophysical parameters and relevant neutrino fluxes are provided. Wide-spread adoption of these recommendations will make it easier to compare and combine future dark matter results.
Introduction
Direct dark matter (DM) detection experiments search for rare interactions between DM particles and nuclei in ultra-quiet underground laboratories. As sensitivity improves, the nuances of data analysis become increasingly critical. Historically, different collaborations have adopted varying conventions for astrophysical parameters and statistical methods. These choices can shift results by significant margins—sometimes by 50% or more—complicating the task of comparing the sensitivity of different detectors or performing global fits. This paper, representing a landmark consensus among members of major collaborations like DAMIC, LZ, XENON, and PandaX, establishes a unified framework to address these discrepancies.
Statistical Inference: The Shift to Two-Sided Tests
A core recommendation is the adoption of the Profile Likelihood Ratio (PLR) method using a two-sided test statistic. Many previous analyses used one-sided tests, which can lead to "flip-flopping"—switching between reporting an upper limit and a discovery based on the observed data.
As shown in Figure 2, the choice between one-sided and two-sided statistics has a measurable impact on the reported cross-section ($\sigma_{SI}$). The two-sided approach provides a continuous transition from setting limits to claiming a discovery, avoiding the statistical pitfalls of switching methods post-hoc and handling cases where background fluctuations might lead to "too-strong" limits.
Standardizing the Astrophysical Environment
To translate event rates into particle physics properties, researchers must model the dark matter halo. The paper updates the Standard Halo Model (SHM) using modern astronomical data from surveys like Gaia and RAVE. The recommended values for the 2020s are:
- Local DM density ($\rho_\chi$): 0.3 GeV/cm³
- Local standard of rest velocity ($v_0$): 238 km/s
- Galactic escape speed ($v_{esc}$): 544 km/s
Furthermore, the paper provides a standardized table for neutrino fluxes. As experiments reach the "neutrino fog," where solar and atmospheric neutrinos mimic DM signals, having a common background model is essential for consistent reporting across the field.
Reporting Results and the Look-Elsewhere Effect
When searching across a broad range of WIMP masses, the probability of seeing a local fluctuation increases—a phenomenon known as the Look-Elsewhere Effect (LEE). The paper recommends reporting "global" p-values that account for this trial factor, using a $3\sigma$ global discrepancy for "evidence" and a $5\sigma$ threshold for "discovery."
Figure 3 illustrates the recommended presentation of results. The top panel shows the exclusion limit ($\sigma_{SI}$) with $\pm 1\sigma$ and $\pm 2\sigma$ sensitivity bands. Crucially, it includes a Power Constrained Limit (PCL) (dashed line) with a threshold of $\pi_{crit} = 0.32$. This prevents experiments from claiming sensitivity in regions where they have no physical ability to detect a signal due to downward background fluctuations. The bottom panel displays the local p-value, providing a clear visual of the significance of any observed excesses against standard thresholds.
Conclusion
By adopting these conventions, the dark matter community moves toward a more transparent and collaborative era. While the Standard Halo Model is a known simplification, its use as a benchmark allows for direct comparison between experiments. These recommendations provide the necessary "standard yardstick" for the next generation of high-sensitivity dark matter searches, ensuring that the field's competitive nature is supported by rigorous, comparable analysis.
Summary updated