Generalised Linear Models SPS5032
- Academic Session: 2019-20
- School: School of Social and Political Sciences
- Credits: 20
- Level: Level 5 (SCQF level 11)
- Typically Offered: Semester 1
- Available to Visiting Students: No
- Available to Erasmus Students: No
This is an advanced course on regression modelling and focuses on the Generalised Linear Model (GLM) and the maximum likelihood principle.
Two hours lecture and one hour tutorial per week.
Expected delivery of lectures Mondays 4-6pm. Tutorial to be placed through the week with sign up on MyCampus.
Requirements of Entry
It is expected that an introductory post-graduate social science statistics course is completed in advance of this course. For example: the College of Social Science Graduate School's Quantitative Data Analysis course or a comparable course.
An excerpt from a population survey is provided to the students. Each student identifies a research question, proposes an advanced statistical model covered in the course, applies it in a methodologically sound way to the dataset using R in order to answer the research question, justifies the analysis in terms of whether the assumptions of the model are met, and interprets the results. A 3,000-word report on the research question, data analysis, diagnostics, interpretation, and conclusions is submitted by each student.
This is an advanced course on regression modelling and focuses on the Generalised Linear Model (GLM) and the maximum likelihood principle. These techniques are frequently employed in contemporary quantitative research and can be found in publications across a range of subjects. The course starts where the course "Quantitative Data Analysis" ends. The linear model is re-interpreted as a special case of the generalised linear model, and other outcome distributions of the GLM are introduced, such as models for binary, ordinal, multinomial, count, and event history data. The maximum likelihood principle is discussed as the GLM's main estimation strategy. Advanced specifications, such as interaction terms, random effects, and robust estimation, are introduced. The main objective of the course is to give students a solid working knowledge of regression modelling for various scenarios that go beyond the standard case of the linear model. Students will learn how to apply and interpret generalised linear models and related techniques and acquire a solid understanding of how to model social phenomena with the tools of statistical inference. The statistical techniques are taught theoretically, through the use of examples, and in the statistical computing environment R.
Intended Learning Outcomes of Course
After taking this course, students should...
•Be able to apply the Generalised Linear Model (GLM) to observational data on social phenomena and transfer the methods to their own research questions.
•Have a solid understanding of the different outcome distributions and techniques corresponding to different types of data-generating processes, including their assumptions and limitations.
•Be able to interpret parameter estimates and uncertainty measures of each technique.
•Understand how the maximum likelihood principle and maximum likelihood estimation work and how they can be applied to empirical data.
•Know how to implement GLMs and GLMMs in the statistical computing environment R.
•Be able to understand and evaluate the statistical analyses in current published articles in highly ranked journals in the social sciences and design and execute similar analyses.
Minimum Requirement for Award of Credits
Students must submit at least 75% by weight of the components (including examinations) of the course's summative assessment.