Microcredential ekomex Coincidence Analysis for Causal Learning

This hands-on three-day online course teaches how to learn complex causal structures from data using Coincidence Analysis (CNA), a configurational method used in health and social sciences for modeling how causes and context factors bundle into alternative paths to outcomes.

What Is This Course About?

This three-day online compact course introduces participants to Coincidence Analysis (CNA), a configurational method for causal data analysis and modeling that is widely used in health and social sciences and uniquely designed to capture causal complexity and context-sensitivity. Many causes do not operate in isolation but only become effective in combination with suitable context factors; moreover, the same cause placed in different contexts can have radically different effects, such that there is no significant pairwise association between causes and effects on the population level. To causally model such structures, CNA bundles multiple factors into complex causes in which every element is indispensable for producing the outcome, and places these bundles on the same or on alternative causal paths to one or more outcomes. The course covers the conceptual foundations, the search algorithm, hands-on data analysis using the R package cna, model evaluation, and strategies for robustness analyses and dealing with model ambiguities; no prior knowledge of CNA is required, though familiarity with R is an asset.

Learning Goals

This course will teach you to:

• Understand the philosophical and technical foundations of Coincidence Analysis and its methodological protocols and workflows
• Identify suitable research questions for CNA
• Prepare data for CNA (including calibration and factor selection) and conduct a complete CNA using the cna R package
• Interpret, evaluate, and compare CNA solutions using measures of fit, robustness checks, and available theory and case knowledge
• Critically assess published CNA applications

Recommended Readings for the Course

• Whitaker, Rachel Garr, Nina Sperber, Michael Baumgartner, Alrik Thiem, Deborah Cragun, Laura Damschroder, Edward J. Miech, Alecia Slade, and Sarah Birken. 2020. “Coincidence Analysis: A New Method for Causal Inference in Implementation Science.” Implementation Science 15(1): 1070. https://doi.org/10.1186/s13012-020-01070-3
• Baumgartner, Michael, and Mathias Ambühl. 2025. “cna: An R Package for Configurational Causal Inference and Modeling.” R Package Vignette, Version 4.0.3. https://cran.r-project.org/web/packages/cna/vignettes/cna.pdf
• Rhodes, Jesse H. 2025. “Laboratories of Democratic Renewal: Explaining Substantial Improvement in the Quality of Democracy in the American States.” Political Science Research and Methods: 1–18. https://doi.org/10.1017/psrm.2025.10051

Assignments for the Course

Two formative in-class assignments (not graded):

(1) a hands-on CNA exercise using provided data and R scripts;

(2) a quiz-style exercise to practice and consolidate core concepts and analytical steps.

Schedule

Course dates: 22–24 February 2027 (Monday–Wednesday)

Platform: Online (Zoom)

In addition, students should read the assigned literature in advance.

Day 1 (22 Feb) — Foundations

• 90’ synchronous lab (Boolean algebra, INUS theory of causation)

• 90’ synchronous lab (inference principles and the CNA algorithm)

• 60’ synchronous lab (data types, measures of fit)

• 60’ office hour

Day 2 (23 Feb) — Application

• 90’ synchronous lab (calibration, factor selection)

• 90’ synchronous lab (hands-on CNA analysis with R)

• 60’ independent small group learning (apply CNA to data in R)

• 60’ office hour

Day 3 (24 Feb) — Advanced Topics

• 90’ synchronous lab (model ambiguities)

• 90’ synchronous lab (overfitting, robustness analysis)

• 60’ independent small group learning (quiz-style exercise)

• 60’ office hour

Who Is Your Instructor?

Michael Baumgartner is a professor of philosophy at the University of Bergen, Norway, and the main developer of the Coincidence Analysis (CNA) method and its accompanying R package. He has published extensively on configurational causal modeling, regularity theories of causation, and Boolean methods in journals such as Sociological Methods & Research, Political Science Research and Methods, British Journal for the Philosophy of Science, and Multivariate Behavioral Research, and has taught CNA workshops worldwide for over a decade. His work on CNA has been applied across the social sciences, health sciences, implementation science, and beyond. Web: https://m-baum.github.io/