Applied Quantitative Methods for Management Research - 2026

Module A: Understanding Causality

 

Course coordinator: Oana Vuculescu

Lecturers: Bart Verwaeren; Harry Garretsen, Steffen Triebel, Sarah Bruhs Sørensen

Workload: 5 ECTS

Administrative assistance: Lisbeth Widahl

Time and place: 

Week 41, Monday – Friday, 5-9 October 2026, all days 9-16

Aarhus BSS, Aarhus University, Universitetsbyen, 8000 Aarhus C. Room: 5 Oct.: 1834-142, 6 Oct.: 1812-117, 7 Oct.: 1812-145, 8-9 Oct.: 1834-238.

Application deadline: 2 September 2026

__________________________________________________________________________

How can you design a study that allows you to make credible causal claims, when randomized experiments are not feasible? How can you address concerns about endogeneity or flawed research design raised by reviewers and colleagues? This PhD course equips students with the tools and strategies to tackle these questions, focusing on the challenges of causal inference in management research. Grounded in counterfactual thinking, the course introduces a range of tools designed to uncover causal relationships, focusing on observational and quasi-experimental methods.

Course objectives:

• Understand key concepts of causality, threats to causal inference and strategies to address them in management research.
• Learn the main techniques used in management research to handle endogeneity and identify causal effects, including regression-based adjustment, panel-data and fixed-effects models, matching, difference-in-differences, and instrumental variables, with a closing look at machine-learning approaches at the frontier.
• Learn how to apply some of these techniques using data from published articles.
• Critically review scientific papers to assess the validity of causal claims and strengths and weaknesses of the chosen methods.

The course progresses from the core principles of causality (day 1) through the workhorse tools for addressing specific challenges in causal inference (days 2–4), and closes with instrumental variables and a look ahead to the research frontier (day 5). Throughout the course, an emphasis is placed on understanding the intuition behind these techniques, seeing them applied in published empirical articles, and replicating them using software packages.

The course is designed to help students critically assess causal claims, identify appropriate techniques given the available data, and contribute effectively to academic discourse in management research.

Day 1: Foundations of Causal Inference

Day 1 sets the conceptual foundation for the week. Building on counterfactual thinking, students formalise what a causal claim is using the potential-outcomes framework and learn to reason about identification: the conditions under which data can, and cannot, recover a causal effect. Directed Acyclic Graphs (DAGs) are introduced as a complementary language for encoding those assumptions and diagnosing confounding, selection, and the problem of bad controls. Randomised controlled trials are discussed as the benchmark against which observational designs are judged.

• Counterfactual thinking, potential outcomes, and what makes a claim causal; defining the estimand (ATE, ATT) and treatment-effect heterogeneity
• Randomised controlled trials as the benchmark for causal inference
• Threats to identification: confounding, omitted-variable bias, selection bias, measurement error, and “bad controls”
• Directed Acyclic Graphs (DAGs) for encoding assumptions and identifying confounding
• A framework for credible inference: from research question to estimand, identification, and estimator
• Hands-on exercises

Day 2: Regression-Based Approaches and Limited Dependent Variable Models

Day 2 introduces regression as an adjustment strategy to mitigate confoundedness. Regression models are foundational tools for causal inference, but their use requires understanding their limitations and extensions. This day focuses on core regression concepts and introduces methods for handling binary and count outcomes through Limited Dependent Variable (LDV) models.

• Linear regression (OLS) refresher
• Control strategies: handling confounders, adding interaction terms, and assessing robustness to omitted variables
• Hands-on exercises
• Introduction to limited dependent variable (LDV) models, e.g. logit and count (Poisson) models – the aim is to grasp the intuition behind them, for settings with categorical or non-continuous dependent variables

Day 3: Panel Data and Matching

Day 3 covers two of the core tools for drawing causal inferences from observational data. Panel data methods exploit repeated observations of the same units to control for stable, unobserved differences between them – a major source of endogeneity in management research. Matching then addresses confounding on observed characteristics, constructing comparable treated and control groups so that like is compared with like. Students learn when each approach is appropriate and practise both on firm-level data.

• Introduction to panel data: cross-sectional versus panel data
• Fixed effects and random effects models: controlling for time-invariant unobserved heterogeneity
• Lagged dependent variables and temporal relationships
• The logic of matching on observable characteristics
• Types of matching: propensity score matching, coarsened exact matching; strengths and weaknesses
• Hands-on exercises

Day 4: Difference-in-Differences (DiD)

Difference-in-Differences (DiD) is the quasi-experimental workhorse of empirical management research. Building on the panel-data foundation from Day 3, this day shows how DiD estimates causal effects by comparing changes over time between treated and control groups. Students learn the logic and the key assumptions, and then engage with the recent reassessment of DiD – why the conventional approach can mislead when units are treated at different times, and the newer estimators developed to address it.

• The logic of difference-in-differences: pre/post and treated/control comparisons, and the parallel-trends assumption
• Model extensions
• Dealing with staggered treatment timing, and why conventional two-way fixed effects can mislead
• Modern DiD estimators and applications in management research, with practical hands-on exercises

Day 5: Instrumental Variables and the Frontier

The final day pairs one more classical identification strategy with a look at where the field is heading. Instrumental variables address endogeneity by isolating variation in the treatment that is unrelated to the outcome’s other determinants; the emphasis here is on the intuition and on recognising a valid instrument, rather than on technical implementation. The day then looks ahead to the research frontier, where machine learning is being integrated into causal inference – to estimate how effects vary across units and to adjust for many potential confounders while preserving valid inference.

• Instrumental variables: exogenous variation and the exclusion restriction; when and why IV is used, and how to find valid instruments (emphasis on intuition)
• From average effects to heterogeneous treatment effects: who benefits, and by how much
• Machine-learning approaches to causal inference: causal forests and double/debiased machine learning
• A look ahead: causal discovery, text and large language models, and combining experimental and observational data
• Hands-on exercises

Assignment
Pre-course: there will be pre-readings to prepare for the course. Moreover, participants are expected to submit a short project description (max 1 page) with a focus on the method section.

Post-course: The focus of the final assignment (4-5 pages) will be on applying the course concepts and methods to the student’s own research project and critically evaluating the credibility of its causal claims, including the identification assumptions involved and how sensitive the conclusions are to them.

Application
Deadline for application: 2 September 2026. Please download and fill in the application form. The application form should be sent by email to: Department of Management, Aarhus BSS, Aarhus University, att. Lisbeth Widahl. Please note that your application is binding.

Fee
External participants (from outside Aarhus University) will have to pay a fee to cover lunch and refreshments. For more information, please contact Lisbeth Widahl. Participants will have to make their own arrangements regarding travel and accommodation.