ST952: Introduction to Statistical Practice

This is a short (10 hour) course designed for beginning Statistics M.Sc. students, who come from a wide range of backgrounds but who typically have little or no experience of using statistics to analyse genuine practical problems in depth.

Aims

The overall aim is to familiarise students with what practising statisticians do, including the appropriate use of computing packages such as "S-Plus". Specific aims are:

1. To review and interpret the principles underlying statistical practice.
2. To introduce S-Plus as a computing environment for statistics.
3. To discuss ideas whose importance, interpretation, and indeed existence, are usually learned only from long experience as a practising statistician.

Objectives

After completing this course, students should be able to do the following:

1. Criticise and make constructive comments on all aspects of a statistical problem, including bias, data quality, patterns and interrelationships in the data, modelling assumptions, and the conclusions that may be drawn.
2. Use S-Plus (including, where appropriate, graphics, statistical models, and user-written functions) to analyse statistical data.
3. Write a detailed report on a small data-set (typically containing fewer than 150 individual values) arising from a complex practical problem.

Prerequisites

Undergraduate courses on Pure Mathematics, Probability and some Statistics.

Syllabus

Background

What is statistics?; data and data analysis; probability, its practical value \& interpretation, random variables/vectors and their distributions; types of investigation.

Mathematical statistics

Probability modelling, likelihood and its interpretation, schools of inference; point and interval estimation, hypothesis testing; linear models and some generalisations.

Statistical analysis

The role of assumptions, model checking, transformations and robustness; design; numerical methods and approximation; practical considerations.

Use of S-Plus

Statistical computing, basic S commands, S resources (inbuilt help, books with S code, "Statlib" etc.)

Assessment

20% by class test, 80% by project (set during Term 1, due in at start of Term 2).

Online Resources

Lecture notes (postscript)
Previous assessments: 1997, 1998, 1999.

Books

Chatfield (1995) "Problem Solving: A Statistician's Guide"

Cox & Snell (1981) "Applied Statistics, Principles and Examples"

Venables & Ripley (1997) "Modern Applied Statistics with S-Plus"

Experience is the name everyone gives to their mistakes.

Oscar Wilde.

Research is what I'm doing when I don't know what I'm doing.

Werner von Braun, (attrib.)