A compact graduate course that teaches doctoral students to think clearly about research design — from formulating precise research questions to connecting theory, data, and analysis into a coherent whole.
What Is This Course About?
Good research requires a research design – a structured plan that specifies how a research question will be answered. Yet graduate training rarely makes that design explicit or teaches students to evaluate it systematically. This course introduces a four-element framework for constructing and evaluating research designs in the quantitative social sciences, covering research questions, theoretical models, data collection strategies, and data analysis strategies. Drawing on recent estimand-centered approaches, it helps doctoral students from across the social and behavioral sciences connect the methods they already know into a unified, reflective practice.
Learning Goals
Understand the four elements of a research design and their logical interdependencies
Formulate precise research questions as causal or descriptive estimands
Use directed acyclic graphs (DAGs) to encode theoretical assumptions and assess identification of descriptive and causal quantities
Evaluate data collection strategies (experimental vs. observational, random vs. nonrandom sampling) in terms of what they identify and for whom
Critically assess data analysis strategies, including the distinction between design-based and model-based inference, estimator properties, and the translation of statistical output into substantive quantities of interest
Apply the framework iteratively to evaluate published research and to develop and improve their own research designs
Recommended Readings for the Course
Blair, Graeme, Coppock, Alexander, & Humphreys, Macartan. (2023). Research Design in the Social Sciences: Declaration, Diagnosis, and Redesign. Princeton University Press. [Chapters 1-3 and 5-9]
Lundberg, I., Johnson, Rebecca & Stewart, Brandon. M. (2021). What is your estimand? Defining the target quantity connects statistical evidence to theory. American Sociological Review, 86(3), 532–565.
Rohrer, Julia M., & Arel-Bundock, Vincent (2026). Models as prediction machines: How to convert confusing coefficients into clear quantities. Advances in Methods and Practices in Psychological Science, 9(2).
Assignments for the Course
This is a compact course (3 days, 12 hours) carrying 2 ECTS points. Requirements for credit:
Attendance (minimum 80% of course activities)
Assignment 1 (Day 1 or 2, in-class, formative, ungraded): Diagnose a published study using the four-element framework — identify the research question/estimand, the theoretical assumptions, the data collection strategy, and the analysis strategy; note where the design is strong and where it is vulnerable.
Assignment 2 (Day 3, in-class, formative, ungraded): Apply the framework to your own research — sketch the four elements of your current or planned project, identify the key identification challenges, and indicate how your data collection and analysis strategies address them.
Both assignments are designed to be completed during the practical sessions and will be discussed collectively. No written submission is required beyond the in-class exercise.
Schedule
The course consists of six 90-minute sessions across three days (two sessions per day). The first session opens with a 45-minute introduction and course overview, after which the schedule remains flexible: time is allocated to topics as needed rather than rigidly partitioned. Practical exercises are integrated into sessions rather than reserved for separate slots. Students should expect a mix of conceptual presentation, group discussion, and hands-on application throughout each day.
Who Is Your Instructor?
Peter Selb is Professor of Research Methodology at the University of Konstanz, where he teaches graduate and undergraduate courses in survey methods, causal inference, research design and statistics. His methodological interest spans measurement, population inference, and causal inference in the quantitative social sciences. His substantive research centers on political behavior and public opinion.
Website: https://www.polver.uni-konstanz.de/cdm/people/faculty/selb/
Peter Selb
Lecturer komex
