Physics / advanced
Inference and evidence from physical data
Analyze experimental measurements through generative models, correlated errors, likelihoods, model checking, and reproducible reporting of physical conclusions.
Your route through physics
Follow the path or begin with a question. Every lesson is open.
Your place, at your pace. Progress is optional and saved only in this browser when available.
Read / explore / reason
Inside this module
Module 01 advanced
Inference and evidence from physical data
Analyze experimental measurements through generative models, correlated errors, likelihoods, model checking, and reproducible reporting of physical conclusions.
- 01
Counting statistics and Poisson likelihoods
Build a count-generating model, infer an efficiency-corrected rate, and propagate independent background fluctuations.
29 min - 02
Weighted fitting and correlated errors
Fit a common physical parameter with a declared covariance matrix and identify uncertainty that averaging cannot reduce.
29 min - 03
Maximum likelihood estimation in physical models
Estimate a lifetime with a normalized sampling density and show how truncation changes the likelihood.
29 min - 04
Bayesian posteriors and prior sensitivity
Update explicitly declared probability distributions over parameters and distinguish posterior uncertainty from repeated-sample variance.
29 min - 05
Confidence coverage and interval construction
Construct repeated-sampling intervals and explain why boundaries and discreteness invalidate automatic symmetric approximations.
29 min - 06
Goodness of fit and posterior predictive checks
Compare observed discrepancies with a correctly generated reference distribution and separate parameter estimation from model criticism.
29 min - 07
Systematic error modeling and nuisance marginalization
Represent calibration uncertainty in a joint model and distinguish profiling from integrating a probability distribution over nuisance parameters.
29 min - 08
Bootstrap resampling and dependent data
Identify the resampling unit, calculate a complete small bootstrap distribution, and preserve cluster or time dependence.
29 min - 09
Blind analysis and researcher degrees of freedom
Protect analysis choices from result-driven feedback and quantify the difference between a local test and a search procedure.
29 min - 10
Reproducible figures, tables and data provenance
Preserve the numerical and procedural information needed to reconstruct a physical result, including units, normalization and correlations.
29 min