Academic Training Lecture For Postgraduate Students

Telling the Truth with Statistics (1/5)

by R. Barlow (Univ. of Manchester, UK )

Europe/Zurich
500/1-001 - Main Auditorium (CERN)

500/1-001 - Main Auditorium

CERN

400
Show room on map
Description
This course of lectures will cover probability, distributions, fitting, errors and confidence levels, for practising High Energy Physicists who need to use Statistical techniques to express their results. Concentrating on these appropriate specialist techniques means that they can be covered in appropriate depth, while assuming only the knowledge and experience of a typical Particle Physicist. The different definitions of probability will be explained, and it will be appear why this basic subject is so controversial; there are several viewpoints and it is important to understand them all, rather than abusing the adherents of different beliefs. Distributions will be covered: the situations they arise in, their useful properties, and the amazing result of the Central Limit Theorem. Fitting a parametrisation to a set of data is one of the most widespread uses of statistics: these are lots of ways of doing this and these will be presented, with discussion of which is appropriate in different circumstances. This then leads to the subject of errors and how to calculate them, in which there is a lot of timid overestimation in general practice, and the whole business of confidence intervals, where recent experimental results (with the signal small or zero) have driven new theoretical ideas.
transparencies
Video in CDS
From the same series
2 3 4 5