[Paper Review] LHC EFT WG Report: Experimental Measurements and Observables

N. F. Castro, K. Cranmer+8 more|arXiv (Cornell University)|Nov 15, 2022

Particle physics theoretical and experimental studies4 citations

TL;DR

This paper serves as a comprehensive guide for experimental measurements and observables in the context of global Standard Model Effective Field Theory (SMEFT) fits at the LHC. It evaluates multiple approaches—matrix-element, machine learning, and traditional differential measurements—to identify optimal observables for EFT sensitivity, emphasizing compatibility with global fits and re-interpretation, while highlighting the importance of covariance, likelihood reporting, and systematic treatment for unbiased constraints on Wilson coefficients.

ABSTRACT

The LHC effective field theory working group gathers members of the LHC experiments and the theory community to provide a framework for the interpretation of LHC data in the context of EFT. In this note we discuss experimental observables and corresponding measurements in analysis of the Higgs, top, and electroweak data at the LHC. We review the relationship between operators and measurements relevant for the interpretation of experimental data in the context of a global SMEFT analysis. One of the goals of ongoing effort is bridging the gap between theory and experimental communities working on EFT, and in particular concerning optimised analyses. This note serves as a guide to experimental measurements and observables leading to EFT fits and establishes good practice, but does not present authoritative guidelines how those measurements should be performed.

Motivation & Objective

• To bridge the gap between experimental and theoretical communities in EFT analyses at the LHC, particularly regarding optimized measurements.
• To evaluate and compare different approaches to experimental observables and measurements for EFT sensitivity, including matrix-element methods and machine learning.
• To establish best practices for reporting measurements—such as covariance matrices, multi-dimensional likelihoods, and fiducial regions—for compatibility with global SMEFT fits.
• To analyze the mapping between experimental measurements and SMEFT operators, quantifying sensitivity and identifying key constraints.
• To support legacy data re-interpretation by promoting standardized, re-usable measurement outputs such as truth-level distributions and re-weighting-friendly formats.

Proposed method

• Systematically reviews experimental approaches: matrix-element-based observables, machine learning-optimized observables, and traditional differential/fiducial cross sections.
• Compares single-step and two-step measurement strategies, evaluating trade-offs in bias, optimality, and re-interpretability.
• Applies Fisher information matrix analysis to quantify relative sensitivity of observables to SMEFT operators in the SMEFiT 2021 analysis.
• Uses fitmaker analysis to examine linear dependencies of measurements on EFT operators, enabling intuitive understanding of individual and marginalized constraints.
• Evaluates the impact of dataset variations on global fit sensitivity, identifying dominant measurements per operator.
• Emphasizes the need for full likelihood reporting and proper treatment of correlated systematics to enable future global combinations and reinterpretation.

Experimental results

Research questions

• RQ1 Which experimental observables are most sensitive to new physics effects in SMEFT, and how do they compare across different analysis techniques?
• RQ2 How do different measurement strategies—matrix-element, machine learning, and fiducial cross sections—impact the quality and reusability of EFT constraints?
• RQ3 What is the relative contribution of individual measurements to global SMEFT constraints, and how can this be quantified using Fisher information or fit variations?
• RQ4 How can experimental measurements be reported to maximize compatibility with global EFT fits and enable re-interpretation by external groups?
• RQ5 To what extent do detector effects, unfolding, and systematic uncertainties affect the reliability of EFT-sensitive observables in global fits?

Key findings

• The Fisher information matrix analysis in the SMEFiT 2021 study reveals that certain Higgs and diboson measurements are particularly sensitive to specific SMEFT operators, especially in the flavor-universal scenario.
• In global SMEFT fits, Higgs measurements significantly improve constraints on Wilson coefficients, with the exclusion of Higgs data leading to substantially weaker bounds—demonstrating their critical role.
• Differential and fiducial cross sections, especially when combined with optimized observables, provide strong sensitivity to EFT effects, particularly when energy-growing or interference effects are exploited.
• The fitmaker analysis shows that individual measurements exhibit linear dependencies on specific operators, enabling intuitive interpretation of constraints and identifying dominant contributions.
• Two-step measurement strategies preserve data for re-interpretation but risk introducing bias or sub-optimality if not carefully designed, especially in unfolding and efficiency modeling.
• Reporting full multi-dimensional likelihoods and covariance matrices is essential for future global fits, though complex due to correlated systematics and weakly constrained parameters.