< Back to search results

Statistical Modelling using General Linear Models (GLMs)

Audience: Academic staff, PG research, Professional staff, Research staff, Research supervisors

Date: Monday 12 October 2026 to Monday 31 May 2027

Times: null

Key details: Statistical Modelling using GLMs Extend simple linear regression to generalised linear models.

Target audience

Open to all Postgraduate Research students or University staff members. Note that this course is not open to students on an undergraduate degree or to students on a taught postgraduate programme.

Qualifications needed

This course presumes knowledge of: • basic statistical concepts, in particular hypothesis testing and simple linear regression. The “Introduction to Statistics” course would be an ideal introduction. • basic commands in R required to undertake an exploratory data analysis, fit simple linear regression models and interpret the R output. The R Essentials and Statistics Essentials courses provide a useful refresher of both these requirements (but are not enough by themselves without some prior knowledge).

Course pre-work

Students are encouraged to install R and RStudio on their laptop before the course starts but help will be provided in the introductory session.

Course information

Statistical modelling using generalised linear models (GLMs) is an online, self-paced course covering extensions to linear models and generalised linear models; these models describe some form of response as a function of one, or more, explanatory variables and are powerful tools in understanding complex systems or predicting an outcome.

 

Aims and objectives

    The topics covered are divided into three main sections:

  1. Introduction and extension of Linear Models
    1. Recap of multiple linear regression
    2. Generalized least squares and maximum likelihood
  2. Introduction to Generalised Linear Models
    1. Poisson response data
    2. Binomial (proportions or binary) response data (also known as logistic regression)
  3. More advanced Generalised Linear Models
    1. Multinomial response data (also known as multiple outcome or multi-category models) for data with unordered or ordered categorical responses

 

The course includes a substantial practical component and the methods covered will be implemented using the statistical programming language R and user-friendly environment RStudio.

 

The whole course is spread over semesters 1 and 2 and there will be a guided programme for self-study. Course material (lecture notes, exercises and computer practicals) will be available on Moodle. The time commitment is anticipated to be about 2-3 hours per week (working through the course notes and associated computer practical).

 

There will be an in-person introductory session to get you started which will likely be in week 5 (to be confirmed nearer the time) and throughout the course there will be online help sessions (the exact times and dates TBC). You are encouraged to attend the introductory session in-person to meet the course tutor and fellow participants. Participation in the help sessions is optional. 

 

Tutors

Professional staff, University of St Andrews

Course provider

St Leonard's Postgraduate CollegeEmail: stlc@st-andrews.ac.uk