Abstract
Searches for New Physics at the LHC are described in the framework
of a statistical test. Making a discovery corresponds to rejecting
the hypothesis that the data are consistent with known background
processes, and in addition one uses the data to test various signal
models and to place limits on their parameters. Recent developments
in both Bayesian and Frequentist techniques for carrying out these
tests are described, including methods for incorporating systematic
uncertainties and the "look-elsewhere effect".