Choose timezone
Your profile timezone:
Cavendish HEP Seminars: Errors-on-Errors and Higher-Order Asymptotics in Particle Physics
→
Europe/London
Ray Dolby Centre Seminar Room B, B2.002
Ray Dolby Centre Seminar Room B, B2.002
,
Description
In particle physics, data are collected and compared to theoretical predictions by constructing a statistical model that captures both the intrinsic stochastic nature of the measurement process and potential biases introduced by approximations in theoretical predictions and in the modeling of the experimental setup, referred to as systematic uncertainties. However, these uncertainties are often derived from limited data or estimated through ad-hoc procedures, introducing uncertainty in the uncertainty itself - informally referred to as ``error-on-error’'.
In this talk a statistical framework to address this issue is presented. This is achieved by explicitly treating systematic uncertainties as imperfect estimates of unknown error parameters, which effectively incorporates errors-on-errors into the statistical model. This formulation naturally reduces sensitivity to outliers and captures internal tensions within datasets by interpreting them as an additional source of uncertainty. After a formal introduction to the model, we illustrate its application to several particle-physics problems, including the determination of parton distribution functions.
The inclusion of errors-on-errors can also limit the accuracy of standard asymptotic approximations, motivating higher-order corrections for goodness-of-fit tests and confidence intervals. This connects naturally to discovery searches with small event yields, where large-sample approximations can likewise become inaccurate. Counting experiments with backgrounds estimated from control measurements provide a simple setting in which to examine these limitations and demonstrate how higher-order methods improve estimates of observed and expected discovery significance.