Last modified: 28 Jun 2018 10:27
This highly regarded course will take your understanding of statistics to the next level and provide you with the skills and confidence to analyse your complex biological data. Through a combination of lectures, computer based practicals and group work you will gain an understanding of how to deal with pervasive issues in the analysis of real world biological data such as heterogeneity of variance and spatial and temporal non-independence. Hands on computer tutorials will allow you to apply statistical models using modern statistical software (R) to real data, collected by researchers to answer real biological questions.
|Session||First Sub Session||Credit Points||7.5 credits (3.75 ECTS credits)|
This course will be divided into themed weeks during which you will gain experience in understanding complex sampling methodologies and dealing with pervasive issues in the analysis of real world biological data. You will be taught using a combination of lectures, computer practicals and directed group work and emphasis will be placed on the practical implementation of various modelling strategies using the statistical programming environment R.
Week 1: Following a recap of linear models, you will be introduced to some the limitations of using standard linear models for analysing biological data and gain experience in identifying common issues arising from model misspecification. During this week you will focus on dealing with the common issue of heterogeneity of variance using a generalised least squares (GLS) approach.
Week 2: During this week you will learn how to fit models which can account for correlated data arising from repeated measurements from the same sampling unit or from sampling units that are not spatially independent. You will learn to extend the GLS approach introduced in week 1 to model this non-independence.
Week 3: The final week will bring together concepts introduced during the first two weeks and introduce you to analysing data from complex experimental or survey designs using the linear mixed effects modelling framework.
The assessment for this course will be based on one independent report (100%) which will test your conceptual understanding of the topics raised throughout the course.
This is the total time spent in lectures, tutorials and other class teaching.