Postgraduate Course: Advanced Statistics (PUHR11116)
|School||Deanery of Molecular, Genetic and Population Health Sciences
||College||College of Medicine and Veterinary Medicine
|Credit level (Normal year taken)||SCQF Level 11 (Postgraduate)
|Course type||Online Distance Learning
||Availability||Not available to visiting students
|Summary||This online course describes a selection of statistical methods to deal with more advanced problems in epidemiology, building on the principles and approaches described in the course Advanced Epidemiology (PUHR11112). It covers methods to assess the accuracy of predictive models, deal with missing data and allow for longitudinal data and other problems that involve correlated measurements.
This online course is designed to allow students who have developed some expertise in statistical modelling to expand the range of methods with which they feel competent, and to learn a general approach to dealing with inferentially-complex situations. It develops ideas discussed in Advanced Epidemiology (PUHR11112), and examines the statistical aspects and implementation of these in the R statistical programming environment.
Three main areas will be covered:
1. Assessing the accuracy of prediction models using diagnostic test measures, Breslow¿s C-index and related methods. It will also introduce the bootstrap for this purpose, as well as discussing its use for solving inferential problems that otherwise have poor or unknown distributional behaviour.
2. Detecting missing data using a mixture of simpler exploratory/ graphical methods, and adjusting models for this issue using multiple imputation.
3. An introduction to the use of mixed effects models, particularly for longitudinal data.
The course will draw out both the unique features and similarities to other types of statistical inference, as well as discussing model diagnostics checking, variable selection and model building. The course will consist of a mix of recorded lectures, readings, and practical exercises conducted in the open source R statistical programming language. For each practical, students will be encouraged to discuss their analysis plans, share useful code shortcuts and discuss their results on the discussion boards. The data sets for these exercises will be based on real epidemiological data sets and/ or common problems, although the principles will be applicable more broadly across a wide range of medical and scientific research.
Entry Requirements (not applicable to Visiting Students)
||Other requirements|| None
Course Delivery Information
|Academic year 2022/23, Not available to visiting students (SS1)
|Learning and Teaching activities (Further Info)
Programme Level Learning and Teaching Hours 2,
Directed Learning and Independent Learning Hours
|Assessment (Further Info)
|Additional Information (Assessment)
||Written Exam 0 %, Coursework 100 %, Practical Exam 0 % «br /»
The assessment will consist of 2 coursework projects.
||Students will receive continuous feedback via Discussion Boards, and undertake a short formative practical exercise around week 4 of the course, which will be in the style of the final assessment.
|No Exam Information
On completion of this course, the student will be able to:
- Apply a range of advanced statistical methods to data analysis problems in epidemiology.
- Critically evaluate a number of complex statistical issues in epidemiological research in order to select appropriate methodological approaches.
- Use the R statistical programming language for several advanced applications, and interpret and communicate output appropriately
- Show autonomy in the design and execution of advanced statistical analysis in epidemiology.
|A reading list will be provided on the course virtual learning environment.|
|Graduate Attributes and Skills
||The skills developed by this course are key for most types of epidemiological enquiry, and thus fall broadly under the overarching Enquiry and Lifelong Learning attribute. In particular, the core tasks of analysis and project work involve problem solving, critical thinking and evaluation, which map closely to the Research and Enquiry cluster. However, this will also foster Personal and Intellectual Autonomy, contributing to the student's ability to conceive, design, execute and interpret epidemiological research.
|Keywords||Statistics,statistical methods,R,mixed effects,bootstrap,missing data,imputation,prediction
|Course organiser||Dr Niall Anderson
Tel: (0131 6)50 3212
|Course secretary||Miss Suzanne Newall
Tel: (0131 6)50 2022