Statistical Methods (2010-11)

This course runs throughout Michaelmas Term. Miss Burke will teach the material on linear models, and the remainder of the course is taught by Dr Tomas.

Course Aims

The aims of the Statistical Methods course are that students:

1. Can correctly apply and interpret a range of techniques for the analysis of data;
2. Are able to recognise the circumstances in which different methods are appropriate, and apply this to real-life situations;
3. Develop a critical awareness of the assumptions and decisions made during an analysis, and the influence of external factors;
4. Are able to combine different sources of information to draw reasonable conclusions from data; and
5. Are able to communicate their methods and results accurately and effectively.

The course description is available in the Student Handbook. Lecture notes and additional materials will be added to this page as they become available.

Lecture Notes

• Summarising Data. D: 2.1; V&R: 5.1-5.3; R&S:1.5.1
• Estimation and Hypothesis testing (updated Oct. 21). D: 3.1, 3.2, 7.3; V&R: 5.4, 5.5; R&S: ch 2, 3, 4, 19.3, 19.4. I suggest you also read Ripley's "notes and references on robustness" (pages 1 to 6), available from the course page from five years ago.
• Simulation (updated Oct. 21). D:3.3. Because we didn't get time in class to go over them, two exercises have been added to the end of the notes. You might also look at the R script used for the examples in the notes. Note that you can safely ignore the code on the Galapogos Island data until we cover log-linear models in a couple of weeks.
• Linear models Part 1 and Linear models Part 2. There is also some additional material on contrasts, orthogonality and polynomials and least squares.
• Linear models lecture 3, lecture 4 and lecture 5.
• Logistic and log-linear models. D: 10.1 - 10.6, V&R: ch 7, R&S: ch 18 - 22
• Design of Investigations. You may also find it helpful to read some additional notes prepared by David Cox when he taught the course five years ago. R&S: ch 23, 24

R practicals

• Week 1 - Plots, summaries and exploratory data analysis
• Week 2 - Tests and simulation
• Week 3 - Linear regression
• Week 4 (Assessed) - Linear Models Worksheet and dataset. A sample report is now availale, along with some relevant R code. Note that there are many ways of writing and formatting the report and carrying out the analysis, and these documents are only illustrative.
• Week 5 - Logistic and log-linear models

Problem Sheets

There will be four problem classes, held at the following times:
Monday week 3 and week 5, 2-3 pm.
Tuesday week 7 and week 9, 3-4 pm.

• Exercises on lecture 1-4 are now available, with accompanying files PermSim.R and yield.R. Your solutions are due by 4 pm Friday week 2. Solutions are here along with some notes from the marker.
• Exercises on linear models are now available - due by 4 pm Friday week 4
• Exercises on methods for count data and GLMs and accompanying data file Atomic.txt - due by 10 am Monday week 7.
  Outline solutions to questions 1,2,3,5 and 6 are available, as is a description of one possible analysis of the atomic data.
• Exercises on Design of Investigations and accompanying solutions

---

Last modified 16 December 2010