Most postgraduate projects require statistical support at some level. As a principle, the ICM expects that your group will make provision for support in the same way that they would for any other expert resource. This is particularly the case where you need esoteric statistical methods or software.
However, there is some capacity for generic support in statistics and research design. You can contact us in the first instance, and we will try to direct your question to the most appropriate person.
We also provide the following documents and links to resources that you may find useful:
Reading materials
These Research Design Course Notes are taken from one of our MScs. They describe some of the basic principles of research design and medical statistics in simple language, notably with no maths. Likely most helpful for those with no or very little experience.
An absolutely fantastic online statistics textbook that covers a lot of ground. I liked it so much that I bought the book. Bear in mind though, this is not an easy read. Best for those who have experience, or who know what statistical methods they are using.
How to upset the statistical referee is an entertaining read from Martin Bland. I believe this should be required reading; you will learn something to make you a better scientist, even if it’s just to avoid plotting your data like this. And if you wonder why dynamite plots are such a bad thing, even more reason to read the article.
Software
First, check what you need isn’t available for free as part of Newcastle University’s IT package. For example: MATLAB; SPSS; SAS; Minitab; GraphPad Prism; SigmaPlot are all available for the duration of your postgraduate degree.
A page of power calculation apps. If you don’t know what a power calculation is, then you might want to stay away from it.
G*Power is a free stand-alone application for power calculations. See previous warning.
R is a free and open source language for statistical programming. It is enormously powerful and there is a huge amount of help available on the internet, but R is not for the faint-hearted. As a minimum, you should have some experience of programming languages – or be prepared to get it.
R studio is an integrated development environment (IDE) for R. It makes R programming more user-friendly than out-of-the-box. But see previous warning.
Python is worth considering as an alternative to R. It doesn’t have the same active community in statistics, BUT it is a modern and better-thought-out language. If you go down the Python route then in addition, NumPy and SciPy are probably relevant.