Regression workshop series

Session 1
Regression - An Introduction with a Focus on the Model

Regression is at the foundation of many advanced statistical modelling methods, so it is important to gain a firm foundation first in simple linear regression. This workshop introduces regression, from the perspective of modern statistical emphasis on modelling rather than null hypothesis testing. Using a worked example, we will step you through the process of specifying a simple linear regression model, and evaluating the fit using summary statistics as well as diagnostics. The focus will be on understanding the inputs and interpreting the outputs. Participants will gain an understanding of the regression modelling cycle that will help equip them to build their own models in future.

We have allowed two hours for this session, however depending on how many questions and how much discussion develops, we may not require the whole time.

Session 2
An Introduction to Multiple Regression with a Focus on the Model

Regression is at the foundation of many advanced statistical modelling methods, so it is important to gain a firm foundation in regression with multiple predictor variables.

This workshop follows the perspective of modern statistical emphasis on modelling, rather than null hypothesis testing and builds on the Introduction to regression workshop, introducing additional variables. This emphasises the strong links between the testing of a research hypothesis and the model, where multiple processes may be affecting the response variable of interest. Using a worked example, we will step you through the process of specifying a multiple linear regression model and guide the interpretation of the outputs. In particular we will look at how to evaluate the fit using summary statistics as well as diagnostics when comparing several models. The focus will be on understanding the inputs and interpreting the outputs. Participants will gain an understanding of how to use multiple variables in regression that will help equip them to build their own models in future.

Session 3
Introduction to Generalised Linear Models

Generalised linear models (GLMs) are an important tool in a range of research fields for analysing data that have many different forms (such as count, categorical, binary data and others).  They are literally a generalisation of the linear regression model so that the same model form can be used with different data types. 

This workshop will introduce GLMs and cover the concepts behind their use to demystify these models and equip participants with the knowledge to identify the appropriate model for their research question and data.   As there are many types of GLMs, a worked example will be used to demonstrate the concepts behind choosing, fitting and evaluation a GLM for one data type.   

Building on the worked example, an overview of several commonly used GLMs will be given to help identify the appropriate model for other data types.

We have allowed three hours for this session, however depending on how many questions and how much discussion develops, we may not require the whole time.