Postgraduate Course: Statistical Modelling in the Social Sciences (PGSP11486)
|School||School of Social and Political Science
||College||College of Humanities and Social Science
|Credit level (Normal year taken)||SCQF Level 11 (Postgraduate)
||Availability||Available to all students
|Summary||This course introduces a set of models that deal with various forms of dependent variables and that may be used for investigation of complex causal relationships, mainly in the context of social sciences. The models are introduced through the statistical learning techniques although links to generalised linear models are also considered. Lectures are combined with practical computer lab tutorials in order to illustrate the applications of the theoretical tools. The analysis is carried out using the statistical software environment R, which is freely available under the GNU General Public License.
The use of quantitative models to investigate a wide variety of social outcomes, to consider the evidence for causal relationships between a set of explanatory variables and a variable of interest are likely to be useful across many social science subjects. This course provides a foundation for other specific advanced quantitative modules such as Longitudinal Data Analysis, Multilevel Modelling and Bayesian Statistics that provide logical extensions.
There are three main objectives of this course:
1. Provide a unifying framework for linear models through statistical learning with reference to the generalized linear models framework.
2. Introduce some common learning algorithms. (Dimensionality reduction techniques such as PCA and factor analysis, clustering algorithms, and discriminant analysis will be discussed.)
3. Introduce specific models to deal with complex causal relationships (IV regression and selection model)
On top of the theoretical tools introduced, the course aims to equip students two other computational skills: data management and data visualization. R packages dplyr and ggplot2 will be introduced and used for these purposes.
Topics typically covered include:
Data Management and Visualization
Generalized Linear Models
Unsupervised Learning (PCA/Explanatory Factor Analysis, Clustering)
Supervised Learning (Discriminant Analysis)
Entry Requirements (not applicable to Visiting Students)
||Other requirements|| None
Information for Visiting Students
|High Demand Course?
Course Delivery Information
|Academic year 2018/19, Available to all students (SV1)
|Learning and Teaching activities (Further Info)
Lecture Hours 32,
Programme Level Learning and Teaching Hours 4,
Directed Learning and Independent Learning Hours
|Assessment (Further Info)
|Additional Information (Assessment)
||Eight weekly exercises (40% in total).
End-of-course practical project (60%)
The end of course assignment differs to the sister course at undergraduate level in that at level 11, students will be expected to undertake and convincingly communicate all aspects of the quantitative research process relating to a research question of their choice. They will be required to identify a secondary data source to apply the techniques covered on the course to address their research question. The project could be related to the MSc dissertation, PhD thesis, or something entirely separate. Informal one-to one meetings with students (or feedback on project plans) will provide additional support for students to undertake this level 11 assignment. In summary the level 11 assignment requires students to demonstrate not only an ability to apply and interpret complex models critically, but also the capacity to link together theory, data preparation, modelling choices and interpretation of statistical output in an intelligent and convincing way.
||Verbal feedback will be given throughout the course.
Written feedback will be given on the assignments.
A completed and annotated version of the test will be published on Learn.
Eight exercises delivered in class using clickers to respond to questions.
|No Exam Information
On completion of this course, the student will be able to:
- Cast social science research questions in an intelligent, critical and creative way using a theoretically informed statistical model and an appropriate social science dataset
- Critically apply a unified conceptual and mathematical understanding of linear models
- Exercise substantial autonomy to use the R software for data management, data analysis and data visualization including the use of syntax to record and reproduce analysis
- Critically analyse multidimensional data through dimension reduction, clustering and discriminant analysis
- Critically evaluate sophisticated statistical issues, including endogeneity and omitted variable bias, and the implications these carry for causal claims in the social sciences.
|Madsen, Henrik, and Poul Thyregod. Introduction to General and Generalized linear models. CRC Press, 2010. |
James, G., Witten, D., Hastie, T., Tibshirani, R. An Introduction to Statistical Learning. Springer. 2013
Matloff, Norman. The Art of R Programming: A tour of statistical software design. No Starch Press, 2011.
Grolemung, F., Wickham Hadley. R For Data Science. 2016
Chang, Winston. R Graphics Cookbook. O'Reilly Media, Inc., 2012.
Madsen, Henrik, and Poul Thyregod. Introduction to General and Generalized Linear Models. CRC Press, 2010
|Graduate Attributes and Skills
||Developing advanced quantitative skills and the capacity to use them in applied scientific context.
Generic cognitive skills relating to the research process (e.g. evaluation, critical analysis).
Communication, numeracy and IT skills.
Autonomy, accountability and working with others.
|Course organiser||Dr Ugur Ozdemir
Tel: (0131 6)50 3990
|Course secretary||Mr Jack Smith
Tel: (0131 6)51 1485