Much of modern International Relations research relies on an analysis of data with a large number of observations. Some would argue that large N studies are the only way one can make general social scientific inferences. While we don’t concur, it is necessary to understand statistics if one is going to engage in much of the IR. literature, and good quantitative analyses can add to any theory testing. Indeed a general understanding of the principles of statistics can aid the qualitative researcher. This module aims to introduce students to quantitative statistical methods by immersing them in some of the most basic (and fundamental) building blocks of quantitative analyses.
At the most basic level, students should be able to read, evaluate and engage with politics and IR. literature based on quantitative techniques. They should also understand and be able to apply the basic concepts of statistics to their research questions. Therefore they could build a database relevant to their question and using powerful statistical software, Stata 9, generate some results of summary and inference.
1. Measurement - scales and indices
This class introduces the basic of measurement. Any statistical analysis
presupposes that the concepts are (and can be) measured reliably and validly;
therefore it is very important to understand issues in measurement.
2. Types of data – nominal, categorical, ordinal and interval
Concepts can be measures differently - some lending themselves better to
some ways than others - and which way these are measured will have an impact
on what one can do with the data. It is therefore important to understand
what way different variables measure different concepts.
3. Introducing datasets for International Relations
This class introduces students to some of the data resources available to
academic social scientists, such as the Correlates of War dataset. Building
on the previous class we will discuss the reliability and validity of certain
indices. In a lab class students will be introduced to and start to use
Stata 9.
4. Data summaries (centre and spread)
One of the main descriptive goals in social science is to portray the nature
of a phenomenon. Basic descriptive statistics enable one to convey a great
deal of information simply and accurately. Using the statistical software
we will start to generate some results.
5. Graphing data
One of the most efficient ways to convey large amounts of information and
data is through graphs. Starting with a graph of the approach and retreat
of Napoleon’s army to and from Moscow, we can see the uses and abuses
of graphing, focussing particularly on the importance of reasonable scaling.
Students will also generate their own graphs using Stata 9.
6. Sampling, estimation and statistical inference
One of the factors that make statistics so powerful is that we can make
reasonable inferences about some sampled data to the whole universe of cases.
We will discuss sampling and inference based on an understanding of the
Central Limit theorem.
7. Hypothesis testing
The other reason statistical analyses are so powerful is their ability to
between distinguish random variation and real trends. This class introduces
the logic behind and operation of t-tests, which enables (for instance)
one to see if two sets of numbers are ‘different’. This is demonstrated
on real data using Stata 9.
8. Model Building – causality and association; correlationsThe validity of any such tests usually depends on reasonable model specification. In order to avoid spurious relationships it is vital that one understands and has a theoretically defensible causal model. Students will be reintroduced to the concepts of functions, and dependent and independent variables.
9. Cross tabulation
Cross tabulation is one of the more common bivariate techniques for analysing
categorical and ordinal data. It enables one to see if there is a relationship
between two (or more) such variables. Students will be shown how to perform
such tests and interpret the statistical output.
10. Regression analysis
Regression analysis is the most common and useful multivariate techniques
which enables to test certain causal models using most types of data. Students
will be shown how to test models using regression analysis and interpret
the results of the analyses.
11. Choosing the right technique
One of the main problems students have in dealing with statistics is knowing
what technique to use in what circumstance. This class will give students
a way to ascertain this, and in doing so will give an overview of some other
techniques not dealt with in this module.
Weekly assignments: 100%
Much of the work will be conducted in the lab or on students own PCs, and the following texts should be very useful for this. However, many more textbooks exist so students can feel free to find your own.
Below are some useful readings which show how quantitative techniques can be used to answer research questions in IR. One of the main areas is the ‘Democratic Peace’ thesis or predicting conflict behaviour according to regime type, but other questions of democratisation, development and institutional design also lend themselves to statistical analyses, as can studies of national identity and indeed any research question one cares to imagine.
