Postgraduate Course: Bayesian Statistics for Social Scientists (PGSP11553)
|School||School of Social and Political Science
||College||College of Arts, Humanities and Social Sciences
|Credit level (Normal year taken)||SCQF Level 11 (Postgraduate)
||Availability||Available to all students
|Summary||This course covers the basics of Bayesian analysis, drawing on a range of examples from across the social sciences. It covers the philosophical reasons for employing Bayesian methods and discusses their limitations. Attention is brought to practical issues related to the estimation of different types of models. Models covered include, comparisons of means (t-tests) and single level regression models for different types of outcome variables, along with overviews of more advanced issues including multilevel modelling, latent variable models, and missing data.
Analysis is undertaken using the R statistical environment and the teaching of all statistical models is rooted in specific social science related examples.
This course provides a practical introduction to the use of Bayesian statistical methods within social science.
The course begins by considering how a Bayesian approach to statistics varies from the more common Frequentist interpretation. As well as exploring the philosophical differences between the traditions, the strengths and weaknesses of each approach is discussed with particular reference to their application across the social sciences.
The course then considers how Bayesian methods can be implemented with regards to different generalised linear models (covering continuous, binary, ordinal and count outcomes). The central focus of this work is on constructing and interpreting models, illustrated by examples relevant to a range of social science disciplines.
The course concludes by considering more advanced statistical models, which are generally not easily estimated without using a Bayesian approach. These include multilevel models, latent variable models and models with missing data.
The course is practical in nature with students fitting models using the open source software packages R and Stan - a familiarity with R syntax for fitting glm models will be useful, but is not assumed.
Entry Requirements (not applicable to Visiting Students)
||Other requirements|| None
Information for Visiting Students
|High Demand Course?
Course Delivery Information
|Academic year 2022/23, Available to all students (SV1)
|Learning and Teaching activities (Further Info)
Seminar/Tutorial Hours 20,
Programme Level Learning and Teaching Hours 4,
Directed Learning and Independent Learning Hours
|Assessment (Further Info)
|Additional Information (Assessment)
||Coursework submissions are as follows:-«br /»
1. A mid-semester guided assignment where students are required to run, and report, some simple analysis relating to specific variables. (25%)«br /»
2. A project report. Student complete a 3250 word project report analyzing a dependent variable of their choice. Students will be required to identify an appropriate dataset and use a non-linear dependent variable This report will include developing a research question, justifying their choice of dataset and chosen statistical method, along with presentation and interpretation of relevant statistical results. An annotated syntax file must be included (not to count as part of the word count) (75%)«br /»
||Students will receive written feedback on the mid-term assignment within 15 working days.
In addition, they will receive verbal feedback on the non-assessed group presentations in weeks 9 and 10 of the course, and PG students will receive a formative exercise on running regression models in R early in the course (to reflect how they will likely have had less exposure to the software than Q-Step UG students).
|No Exam Information
On completion of this course, the student will be able to:
- Be able to critically discuss the principles and philosophical approach of Bayesian analysis;
- Be able to identify appropriate modelling strategies for a range of social science related research questions;
- Be able to execute Bayesian analysis in an appropriate software package;
- Be able to understand and report Bayesian model output in a way which will be appropriate to both technical and non-technical audiences .
|Kaplan, D (2014). Bayesian Statistics for the Social Sciences. Guildford Press: New York (this would be the nearest to a course textbook for the course).|
McElreath, R (2016). Statistical Rethinking - A Bayesian Course with Examples in R and Stan. CRC Press: London
Kruschke, J (2015). Doing Bayesian Data Analysis. 2nd Edition. Academic Press: London.
Jackman, S (2009). Bayesian Statistics for the Social Sciences. John Wiley: Chichester
|Graduate Attributes and Skills
||Plan and carry out a research project and report its findings appropriately (Skills and abilities in research and enquiry)
Be independent learners who take responsibility for their own learning and are committed to continuous reflection, self-evaluation and self-improvement (Skills and abilities in personal and intellectual autonomy)
Make effective use of oral, written and visual means to critique, negotiate, create and communicate understanding (Skills and abilities in communication)
Use visualisation of quantitative data to convey complex evidence to a general audience (Skills and abilities in communication)
Be able to work effectively with others, capitalising on their different thinking, experience and skills (Skills and abilities in personal effectiveness)
Use of Statistical Package for the Social Sciences (SPSS) and other data processing software and word processing packages (Technical/Practical Skills)
Presentation skills and using presentation software (Technical/Practical Skills)
|Course organiser||Dr Paul Norris
Tel: (0131 6)50 3922
|Course secretary||Mrs Beth Richardson-Mills
Tel: (0131 6)51 1659