Case studies have a long tradition in the social sciences. They have been subject to a great deal of both methodological appraisal and criticism. This course addresses the case study method from a comprehensive perspective and focuses on its methodological and practical dimensions. The course is useful for participants at every stage of their research. Participants at the beginning are provided with the skills to carefully plan their study. Those participants that are in the midst of their analysis can evaluate their case study in the light of what they learn and plan the next steps so as to meet the standards of good case study research.
In the opening session on day 1, we introduce several dimensions that are central to all case study analyses and are important to understand. First, the research goal of a case study can be the development of new hypotheses, the test of hypotheses, and the modification of existing hypotheses in order to solve a puzzle. Second, the level of analysis can be the cross-case level (often understood as the macro level), or the within-case level (i.e. the micro level), or both at the same time. Third, we introduce the importance of difference-making and counterfactuals for causal inference and elaborate the difference between frequentist and Bayesian causal reasoning.
On day 2, we proceed with a basic discussion of causation and causal inference and when and how we claim that an observed empirical association reflects a causal relationship. We introduce the the criterion of difference-making as the benchmark for inferring causal relationships. We further elaborate on the distinction between causal effects and causal mechanisms and their role in causal inference and case studies. In addition, a thorough treatment of causal inference requires a consideration of different notions of causal effects. A distinction is made between correlations (e.g., the more X, the more Y) and set-relations (e.g., if X, then Y) as these currently represent the two major perspectives on causal effects in the social sciences. Finally, we introduce the basics of Bayesianism, as it became increasingly important in the recent literature about case studies and process tracing. As will be seen in the subsequent sessions, the means of implementing a case study may significantly depend on the causal effect that one deems to be in place. One major goal of this session is to encourage the participants to reflect on the specific type of causal relationship they believe to be in place and to formulate their theoretical expectations accordingly.
On day 3, we discuss the importance of delineating the population underlying your case study (provided you aim to generalize). You will see that the specification of the population requires three elements: the definition of scope conditions guiding your analysis, and the positive and negative conceptualizations of the phenomenon that you want to explain in your case study. We start with the distinction between different types of scope conditions and their implications for empirical research. Then, we clarify the distinction between the positive and negative outcome and its role for drawing the boundaries of the population
Day 4 and day 5 are concerned with two related issues. First, we discuss different types of cases and their suitability for answering different research questions. The types of cases that are defined and illustrated include types such as the typical case, the deviant case, and the most-likely case. Second, we consider case selection strategies for each type of case. A discussion of both issues helps the participants to determine what type of study and case selection strategy they need in order to find answers to their research question. Since there is a menu of types and case selection strategies and both are linked to different modes of causal inference (see day 2), we spend two days on these topics.
The second week starts on day 6 with a session on the basics of comparative case studies, including a consideration of John Stuart Mill’s (in)famous method of difference and method of agreement. The session focuses on the benefits and limitations of cross-case comparisons in the light of the existing arguments for and against such designs. These criticisms can be subsumed under the rubric ‘small-n problem’, stating that the number of cases is too small in case studies so as to generate valid causal arguments. This session enables the participants to understand the construction of proper comparisons and avoid making common mistakes in comparative case studies.
On day 7, we continue with a discussion of advanced issues in comparisons. We detail multiple strategies of mitigating the problems that we learned about on day 6. These strategies include the realization of within-unit and longitudinal comparisons, an increase in the number of cases, binary measurement of causes (as opposed to multi-categorical measurement), and the transformation of causes into scope conditions.
Day 8 proceeds with an elaboration of the basics of within-case analysis and process tracing in particular. We relate the idea of a causal mechanism to process tracing, discuss different variants of process tracing and within-case analysis, and how process tracing might help in diminishing some of the problems that one confronts in comparative case studies.
On day 9, we continue with process tracing from a more practical perspective. Different types of sources are introduced (primary sources, interviews, etc.) with a focus on their respective advantages and disadvantages. The goal is to heighten awareness of the problems of fact-gathering and the need to handle the collected information with caution. Based on the discussion of sources and evidence and types of cases (see day 4 and day 5), we elaborate on how to systematically use evidence for causal inference. Specific attention is paid to the logic of Bayesian causal inference (see day 3) and we introduce a step-by-step procedure for making sound causal inferences based on observations.
Finally, day 10 covers the generalization of causal inferences. The opportunities and limits for generalization are discussed in combination with techniques for extending insights beyond the cases under scrutiny. Since case study researchers are often criticized for engaging in generalization, it is important to put these problems in the forefront. In addition, we have a Q&A during which you will have the opportunity to ask all questions that came up during the course and are still open at the end.
SUMMARY: WHAT THE COURSE IS AND IS NOT ABOUT
The course is useful for participants who:
- Want to generate causal inferences for one or multiple cases. Whether or not you intend to generalize your inferences is of secondary importance because this is only one element of the research process (and this course).
- If you eschew causal terminology because you are doing interpretivist, post-structuralist, etc. case studies, you may better attend one of the many ECPR courses covering these philosophies of science. In case you are unsure, please get in touch with me.
- Want to learn the case study method and the careful construction of case study designs. The course has a strong practical element because you will learn tricks and clues for avoiding common mistakes in empirical research, i.e., the practice of case studies. But:
- We do not go into the details of collecting evidence. We discuss the handling of evidence from a methodological perspective, which has some practical implications. But if you only want to learn how to organize and conduct interviews, transcribe recordings, or undertake archival research, you are better advised to take the ECPR interview course (in the case of interviews).
All methodological discussions will be supplemented with examples from case studies from different subfields in political science. Participants are expected to carefully read the obligatory readings and to contribute to the debate in class. For credits (see short course outline above), the exam works as follows: in advance of the Summer School, I will give you three published case studies (journal articles), you have to choose one of the texts and answer five method-related questions (including several subquestions) on the text.
In order to get 8 credits, you must do the small assignments that are given to the participants. This is less work than it might read now because many assignments are based on discussions and work that we will do in class already and refer to your own ongoing research project. In addition, you must submit a research proposal of 10-15 pages after the course in which you, based on the lab parts during the course, apply the lessons of the case study course to your research project (detail your concept, justify your case selection, your cross-case comparison, etc.). Those participants without a research project who want to get ECTS points will be given an equivalent task. The paper has to be sent to me within four weeks (approx.) of the end of the course (exact deadline to be determined by the ECPR).
Before the course starts, participants are invited to fill in a short questionnaire on their research project, their prior methodological expertise, and their expectations regarding the course.