STV4020B - Statistical Analysis
Course description
Course content
Political scientists often pose questions that necessitate analysis of data with particular “complicating” features. Examples are: How does EU membership affect foreign direct investment in the short- and long-term (which requires repeated observations for different countries over time)? Do particular parliamentary institutions affect the number of article changes made to a government bill during the legislative process (which requires count data on number of changes to a particular bill)? Does engaging in international conflict increase or decrease the survival chances of autocratic regimes (which require data on the duration of autocratic regimes)? This course provides a set of tools suitable for analyzing this kind of data and questions.
The course is split in three. The first part focuses on time series and panel data. The second part deals with models that are commonly used for analyzing qualitative or countable dependent variables. The third part introduces survival models, suitable for analyzing the duration of different phenomena. In addition, the course addresses how to interpret and present results to the general audience.
Learning outcome
Knowledge
Students will:
- know different types of data structures and types of dependent variables, including time-series data, panel data, count data, censored data, and duration data
- know how to address issues of autocorrelation in time series and panel data
- know the Random Effects and Fixed Effects models, and their benefits and drawbacks in different situations
- know of autoregressive and first-differenced models, and how to employ the time dimension to better address issues of (reverse) causality, e.g. through lagging or through conducting simple Granger tests
- know how to estimate, evaluate and graphically present results from interaction effects
- know how to use simple simulation tools to interpret and present results
- know how to properly estimate, evaluate and interpret results from various models of discrete choices and outcomes
- know the particular challenges associated with analyzing limited dependent variables and count outcomes
- know the Poisson and Negative Binomial models, and their benefits and drawbacks in different situations
- know the challenges associated with statistically analyzing the duration of events
- know how to transform standard (e.g. country-year) data sets into a duration format
- know simple parametric and the standard Cox semi-parametric duration/survival models
- obtain a good grasp of the benefits and drawbacks various statistical models, and in which situations the different models are more and less appropriate;
- learn how to properly present statistical material visually in tables and figures.
Skills
Students will:
- be able to critically read and evaluate existing statistical studies in political science;
- be able to handle data sets using R;
- be able to apply the various statistical models mentioned above, and learn how to properly test hypotheses, interpret results, and draw careful conclusions;
- be able to replicate statistical studies in political science, and to conduct independent statistical studies.
Competences
Students will:
- enhance their capabilities in carrying out thorough, independent and critical analysis of complex questions;
- enhance their capabilities in critically evaluating empirical research
- enhance their understanding of various elements of the scientific process, including aspects of the relation between theory and empirical evidence and between concepts and measures.
Admission
The course is reserved for students in the Master's programme in Political Science.
Students admitted to other Master's programmes may apply for admission to single master's courses.
Prerequisites
Formal prerequisite knowledge
STV4020A - Forskningsmetode og statistikk or courses with eqiuvalent learning outcomes in OLS regression, logistic regression and a statistical package (such as SPSS or R).
Overlapping courses
6 credits overlap with STV4025 - Quantitative political science
Teaching
10 lectures will be given. The lectures are held intensively for a period of 5 weeks.
Examination
Weekly assignments.
Students will be evaluated on the basis of the weekly assignments. There will not be a final school exam. All assignments must be handed in on time in order to pass the course. Each assignment will be based on a published article and a dataset. The task will be to replicate the study and extend it by using the techniques demonstrated in the lectures. The students need to submit a fully functional script, and a text file of maximum 1000 words that describes the purpose of the study, the method(s) and the results. The deadline for the assignments will be before the first lecture of the following week.
Language of examination
You submit your response in English or Norwegian.
Grading scale
Grades are awarded on a scale from A to F, where A is the best grade and F is a fail. Read more about the grading system.
Explanations and appeals
- Explanation of grades - deadline and processing time
- Appeal about grades
- Complaint about formal exam errors
Resit an examination
Withdrawal from an examination
If you wish to withdraw from the exam you must do so in StudentWeb at least two weeks prior to the deadline. Failure to do so will be counted as one of the three opportunities to sit the exam.
Special examination arrangements
If you have a disability or a health problem that entails significant inconvenience in an examination situation, you may be considered for special examination arrangements. Mothers who are breastfeeding may apply for extra time to complete the exam.
Evaluation
Feedback from our students is essential to us in our efforts to ensure and further improve the high quality of our programmes and courses. All courses are subject to continuous evaluation. At regular intervals we also ask students on a particular course to participate in a more comprehensive, periodic evaluation of this course.