6-10 September 2021
Due to the COVID-19 outbreak, this course will be held online
Introductory statistics are typically taught as a sequence of disconnected tests and protocols (e.g. t-test, ANOVA, ANCOVA, regression) while, in reality, all these analyses can be seen as special cases of a more general linear model. In this course, we will introduce Generalised Linear Models as a unified, coherent, and easily extendable framework for the analysis of many different types of data, including Normal (Gaussian), binary, and discrete (count) responses, and both categorical (factors) and continuous predictors.
The course is aimed at at graduate students and researchers with little statistical knowledge but willing to learn how to extract knowledge from data using statistical models, how to use statistical models to increase our understanding and make predictions about natural phenomena, and acquiring a toolbox to analyse many different types of data (beyond the typical Gaussian responses) using R.
No prior statistical knowledge is required, although some basic knowledge about descriptive statistics and simple linear regression (y = a + bx) is recommended as a good starting point for the course. Students should be familiar with Rstudio and have some fluency in programming R code, including being able to import, manipulate (e.g. modify variables) and visualise data. There will be a mix of lectures, in-class discussion, and hands-on practical exercises along the course.
1. Being able to fit, understand, and use statistical models to make predictions and extract knowledge from data
2. Learn how to analyse data with different statistical distributions, estimating the effects of both categorical (factors) and continuous predictors
3. Visualise data and fitted models to check assumptions, communicate results, and increase understanding
4. Practise R programming, particularly applied to data visualisation and analysis (statistical modelling)
5. Acquire the statistical knowledge required to move on to more complex models (e.g. Generalised Linear Mixed Models, Generalised Additive Models, Bayesian modelling) in the future.
Sessions from 15:00 to 20:00 (Monday to Thursday), 15:00 to 19:00 on Friday (Berlin time). From Tuesday to Friday, the first hour will be dedicated to Q&A and tutorials on practical exercises or students’ own analyses over Slack. Sessions will continuously mix lectures with in-class discussion and practical exercises.
Brief overview of R programming: importing, manipulating, and visualising data.
Introductory statistics. Sampling. P-values, confidence intervals, and significance: overcoming misinterpretations.
Introduction to linear models: mathematical and graphical interpretation. Parameter estimates. Residual variation. Maximum likelihood.
General linear model for Normal (Gaussian) response data. Fitting, understanding, visualising and checking linear models with continuous and categorical predictors. Additive models vs Interactions. Model selection. Making predictions.
Generalised linear models for binary responses. The structure of generalised linear models (GLM). The link function. Fitting, understanding, visualising and checking GLM for binary responses with continuous and categorical predictors.
Generalised linear models for discrete (count) data. Fitting, understanding, visualising and checking GLM for count data with continuous and categorical predictors.
What lies ahead: generalised additive models, multilevel or mixed-effects models to account for hierarchical and correlated data (time, space, phylogeny), Bayesian modelling.
> 30 days before the start date = 30% cancellation fee
< 30 days before the start date= No Refund.
Physalia-courses cannot be held responsible for any travel fees, accommodation or other expenses incurred to you as a result of the cancellation.