Statistics Committee

Feldman-Cousins intervals for binomial errorsThe CMS Statistics Committee forms recommendations on statistical issues identified in CMS physics analyses. Such issues include, but are not limited to, the fitting of theoretical models to data distributions, the estimation of the significance of and limits on various signals, the handling of systematic errors and the combination of results from different analyses.

The committee also communicates with other experiments in order to recommend common methods -- across experiments -- to be used in the extraction of conclusions from the data and also in the combination of CMS results with other external measurements.

The recommendations of the CMS statistics committee appear as written documentation (CMS Notes and/or Web pages) and/or software packages -- some of which may be in common with other experiments.

The members of the committee are expected to act as consultants to CMS collaborators seeking input on specialized statistics issues. Upon the request of a physics group or the physics coordinator, the CMS Statistics Committee also makes formal recommendations (i.e. recommendations that are agreed upon by the committee members and are addressed explicitly to the proponents of an analysis and to physics coordinators) regarding the explicit handling of statistics issues on any physics analysis in CMS.

The committee consists of people with broad and extensive knowledge of statistics issues and is chaired by one person who is appointed for two years. The committee makes its recommendations to the physics group conveners or the physics coordinator, depending on the analysis in question.