The formula for a total cross section measurement is sigma = N_obs / (epsilon * L), where N_obs is the background-subtracted event count, epsilon is the overall efficiency (trigger * reconstruction * selection), and L is the integrated luminosity. Which of these quantities typically carries the largest systematic uncertainty at the LHC?
AN_obs, because counting is imprecise
BThe answer depends on the process: for inclusive W/Z production, the luminosity uncertainty (~1-2%) dominates; for rare processes with complex final states, the efficiency uncertainty (from lepton identification, jet calibration, trigger efficiency) often dominates; and for processes with large backgrounds, the background subtraction uncertainty may dominate
Cepsilon, because detectors are unreliable
DL, because luminosity is always the dominant uncertainty
The luminosity is measured by dedicated detectors (LUCID, BCM in ATLAS; PLT, HF in CMS) calibrated using van der Meer scans, achieving ~1-2% precision. The efficiency is measured using tag-and-probe methods on standard candle processes. For well-measured processes like Z -> ll, the luminosity uncertainty dominates. For top quark pair production, jet energy scale and b-tagging efficiency are comparable. For searches with few events, the statistical uncertainty dominates everything.
Question 2 Short Answer
A 'fiducial' cross section is measured within a restricted kinematic region defined by the experimental acceptance (e.g., lepton p_T > 25 GeV, |eta| < 2.5). Why is a fiducial measurement preferred over a total cross section measurement for comparison with theory?
Think about your answer, then reveal below.
Model answer: A total cross section measurement requires extrapolating from the measured fiducial region to the full phase space using Monte Carlo simulation. This extrapolation introduces model dependence: different generators may predict different fractions of events outside the acceptance. A fiducial measurement avoids this by reporting the cross section only within the experimentally accessible region, where detector effects are well-understood. Theorists can then calculate the predicted cross section in the same fiducial region and compare directly. This approach minimizes model dependence and provides the cleanest comparison between data and theory.
Modern LHC measurements increasingly report fiducial and differential cross sections rather than total cross sections. This philosophy ('measure what you measure, don't extrapolate') has been widely adopted because it provides the most transparent and reproducible results.
Question 3 Multiple Choice
Differential cross section measurements (d sigma/d p_T, d sigma/d y, etc.) are presented as 'unfolded' distributions that correct for detector resolution and efficiency effects. The standard unfolding methods include matrix inversion, iterative Bayesian unfolding, and SVD regularization. What problem do these methods solve?
AThey remove statistical fluctuations from the data
BThey correct for the fact that the measured (detector-level) distribution differs from the true (particle-level) distribution due to finite detector resolution (events migrate between bins), limited efficiency (some events are lost), and background contamination — unfolding inverts the response matrix to recover the true distribution
CThey combine data from multiple experiments
DThey extrapolate the data to higher energies
The detector smears the true distribution through a response matrix R: N_measured = R * N_true + N_background. Simply inverting R amplifies statistical fluctuations (the problem is mathematically ill-conditioned), so regularization techniques are needed to suppress unphysical oscillations while preserving the physics content. Iterative Bayesian unfolding (D'Agostini method) and SVD regularization with a tuneable regularization parameter are the most common approaches. The result is a corrected distribution at 'particle level' (stable particles before detector interaction) that can be directly compared with theory predictions.