ST3ASM: Advanced Statistical Modelling

Module code: ST3ASM

Module provider: Mathematics and Statistics; School of Mathematical, Physical and Computational Sciences

Credits: 20

Level: Level 3 (Honours)

When you'll be taught: Semester 2

Module convenor: Professor Sue Todd, email: s.c.todd@reading.ac.uk

Module co-convenor: Dr Fazil Baksh, email: m.f.baksh@reading.ac.uk

Pre-requisite module(s): BEFORE TAKING THIS MODULE YOU MUST TAKE ST1PS AND ( TAKE ST2LM OR TAKE ST2LMD ) (Compulsory)

Co-requisite module(s):

Pre-requisite or Co-requisite module(s):

Module(s) excluded:

Placement information: NA

Academic year: 2024/5

Available to visiting students: Yes

Talis reading list: Yes

Last updated: 21 May 2024

Overview

Module aims and purpose

In a number of real-life situations the statistical and supervised machine learning techniques described in the second year module on Linear Models are inappropriate. The aim of this module is to describe alternative models that are used instead in many areas of application, to show how they can be fitted, and to indicate how their adequacy can be assessed. Generalised linear models allow us to model non-normal response variables, and so these models are considered in some detail. The module also deals with repeated measurement data, covering both traditional and more modern approaches to analysing such data. This module will enable students to develop an understanding of situations in which different models are likely to be appropriate, illustrating these through a range of examples. Sufficient theory is covered to give students a grounding in the important statistical concepts underlying the models and experience of software to fit the various models will be gained.

Module learning outcomes

By the end of the module, it is expected that students will be able to:

1. Recognise which models should be considered as likely to be appropriate for data arising from different situations
2. Fit appropriate models to data of different types using statistical software
3. Check the adequacy of models, compare alternative models and interpret the results

Module content

The module begins with a consideration of generalised linear models. Features of the generalised linear are described, model fitting techniques are covered and the concepts of deviance and model checking, including overdispersion, are outlined. The linear logistic model for binary (success / failure) data is explored in detail, as are log linear models for data in the form of counts, with a particular focus on analysis of contingency tables. Both models are commonly used in classification problems. For repeated measures data, traditional statistical methods used in the analysis of this form of data will be described, such as the summary statistics approach, split-plot analysis of variance and repeated measures multivariate analysis of variance. More modern approaches utilise mixed models will be considered in detail. 

Structure

Teaching and learning methods

The material is delivered via lectures supported by tutorials which tackle non-assessed exercises, together with computer based practical work. 

Study hours

At least 55 hours of scheduled teaching and learning activities will be delivered in person, with the remaining hours for scheduled and self-scheduled teaching and learning activities delivered either in person or online. You will receive further details about how these hours will be delivered before the start of the module.

Please note the independent study hours above are notional numbers of hours; each student will approach studying in different ways. We would advise you to reflect on your learning and the number of hours you are allocating to these tasks.

Semester 1 The hours in this column may include hours during the Christmas holiday period.

Semester 2 The hours in this column may include hours during the Easter holiday period.

Summer The hours in this column will take place during the summer holidays and may be at the start and/or end of the module.

Assessment

Requirements for a pass

Students need to achieve an overall module mark of 40% to pass this module.

Summative assessment

Penalties for late submission of summative assessment

The Support Centres will apply the following penalties for work submitted late:

Assessments with numerical marks

• where the piece of work is submitted after the original deadline (or any formally agreed extension to the deadline): 10% of the total marks available for that piece of work will be deducted from the mark for each working day (or part thereof) following the deadline up to a total of three working days;
• the mark awarded due to the imposition of the penalty shall not fall below the threshold pass mark, namely 40% in the case of modules at Levels 4-6 (i.e. undergraduate modules for Parts 1-3) and 50% in the case of Level 7 modules offered as part of an Integrated Masters or taught postgraduate degree programme;
• where the piece of work is awarded a mark below the threshold pass mark prior to any penalty being imposed, and is submitted up to three working days after the original deadline (or any formally agreed extension to the deadline), no penalty shall be imposed;
• where the piece of work is submitted more than three working days after the original deadline (or any formally agreed extension to the deadline): a mark of zero will be recorded.

Assessments marked Pass/Fail

• where the piece of work is submitted within three working days of the deadline (or any formally agreed extension of the deadline): no penalty will be applied;
• where the piece of work is submitted more than three working days after the original deadline (or any formally agreed extension of the deadline): a grade of Fail will be awarded.

The University policy statement on penalties for late submission can be found at: /cqsd/-/media/project/functions/cqsd/documents/qap/penaltiesforlatesubmission.pdf

You are strongly advised to ensure that coursework is submitted by the relevant deadline. You should note that it is advisable to submit work in an unfinished state rather than to fail to submit any work.

Formative assessment

Formative assessment is any task or activity which creates feedback (or feedforward) for you about your learning, but which does not contribute towards your overall module mark.

Non-assessed problem sheets and PC class practicals 

Reassessment

Additional costs