We will run four virtual courses for Spring School 2021.
Please note, you can only choose and attend one course option, as they will run alongside each other.
The four main courses will be:
- Causal inference
- Qualitative Methods and Fieldwork
- Machine learning
- Quantitative Text Analysis for the Social Sciences
Each of the main courses will run from 2pm-5pm (GMT), and there will be the option to attend supplementary courses and events at other times throughout the week.
*** Please note, this year's courses will be taught virtually, online.***
Important: All students attending the Oxford Spring School are expected to have access to a computer/laptop with a working installation of R and RStudio with the tidyverse package installed. We have limited capacity to help students overcome problems with hardware/software during the week. You don’t need to be an R expert when you show up, but you need to have the tools to become one.
Instructor: Scott Cunningham (Baylor)
Many times, events happen around the same time and we wonder — was that a coincidence? Or was that something more like causality? What is causality and what are mere coincidences and how can we learn to tell the difference? Cunningham’s course on causal inference lays out the foundational theory of causality rooted in counterfactual reasoning from which economists, statisticians, epidemiologists and more have devised original tools to estimate causal effects both within experiments but as importantly when experiments are impossible. This course will take the student through these contemporary causal designs — matching, instrumental variables, regression discontinuity, difference in differences and synthetic control — and their best estimation methodologies. Students will exit the class with greater comprehension of these methods as well as their practical implementation using R.
Instructor: Jody Laporte (Oxford)
This course introduces students to the logic of inference, concepts, techniques, and data used with qualitative methods in the social sciences. It covers themes like process tracing, small-N comparisons, and within-case analyses. By the end of the course, students will be able to contrast the logic of inference in quantitative and qualitative research and examine their suitability for different research objectives, define and explain concepts like causal chains, causal mechanisms, and causal process observations and identify them in empirical studies, relate causal relationships and mechanisms to the periodization of the analysis, categorize and appraise strategies to select cases in small-N research, formulate standards of assessment to evaluate competing explanations when using qualitative methods, and, finally, evaluate different types of data for their probative value in process tracing.
Instructor: Thomas Robinson (Durham)
Computational techniques broadly referred to as "machine learning” (ML) are attracting increasing attention in the social sciences. This course will introduce students to the fundamental theories and concepts used across ML techniques (e.g maximum likelihood estimation, classification, overfitting, and cross-validation), and how to apply various popular ML techniques (e.g. LASSO, random forests, bayesian additive regression trees, and neural networks). This course will also pay particular attention to inferential problems regarding how best to use ML methods in social science contexts and discuss practical issues (e.g. computational power and parallelisation). Students on this course are required to be familiar with probability and statistics (OLS, hypothesis testing, logistic regression) and to have basic knowledge of the R programming language.
Instructor: Tom O’Grady (UCL)
Texts such as political speeches, social media posts, newspaper articles and international treaties convey a wide range of important information for social scientists. They are now available in digitised form on an unprecedented scale, allowing us to turn texts into data and study them systematically. This course will teach students to collect and analyse texts using statistical and computational methods. By the end of the course they will be able to conduct original research with text as data. The course begins by covering the major approaches to processing and describing documents prior to analysis. It then covers statistical methods that describe, classify and scale documents in order to infer characteristics or traits of their authors, such as partisanship, sentiment and political opinions. Methods covered include dictionary methods, machine-learning classifiers, scaling methods such as Wordscores and WordFish, and topic models. It finishes with the basics of web-scraping for automated text collection online.