IQMR 2026

Integrating Qualitative Methods and Natural Experiments (Modules 17, 21)

Monday, June 22; Tuesday, June 23

Eggers Hall, Room 032

Chris Carter (University of Virginia), Guadalupe Tuñón (Princeton University)

Natural experiments are now a staple of political science research, but their credibility hinges on assumptions that quantitative tools alone cannot evaluate. This module shows where qualitative methods are essential: identifying plausible natural experiments, probing the as-if random assumption, and interpreting what estimates actually mean. We work through instrumental variables, true natural experiments, and regression-discontinuity designs using classic and recent applications, with a shared framework for assessing the credibility of causal claims. We compare these strategies to other methods, including difference-in-differences, most similar case designs, and selection on observables. The goal is practical: sharper instincts for spotting natural experiments in your own work and judging them in others’.

Participants may enter the module sequence after it has begun.

Integrating Qualitative Methods and Natural Experiments I (M17, June 22)

8:45am - 10:15am – Qualitative Approaches to Natural Experiments

We emphasize the crucial role of qualitative methods in identifying and analyzing natural experiments. Through illustrative examples, we demonstrate how qualitative evidence enhances the credibility of causal assumptions and aids in interpreting quantitative results. We also explore how qualitative methods address common criticisms of natural experiments and how natural experiments, in turn, strengthen inferences drawn from qualitative evidence.

Required reading:

Suggested readings:

1:30pm - 3:00pm – Problems in Natural Experiments: Non-Compliance

In this session, we discuss cases where units (e.g., individuals, municipalities, political parties) were assigned to one treatment condition but ultimately receive a different one. We address how these problems arise and can be addressed within the framework of natural experiments. We discuss qualitative methods for identifying non-compliance issues and addressing them.

Required readings:

Suggested readings:

3:30pm - 5:00pm – Problems in Natural Experiments: No Clear and Obvious Random Assignment

One of the most important assumptions of natural experiments is as-if random assignment to treatment. In this session, we focus on cases where there is no clear and obvious randomization procedure. We focus on a sub-class of natural experiments where treatment is assigned based on a unit’s position just above or below a threshold. We discuss the role of qualitative and quantitative methods in understanding the process that determines treatment assignment. We discuss recent applications of these designs and how they might be bolstered by additional qualitative evidence.

Required readings:

Suggested readings:

Integrating Qualitative Methods and Natural Experiments II (M21, June 23)

8:45am - 10:15am – Problems in Natural Experiments: Spillovers

In addition to as-if random assignment to treatment, a second major assumption of natural experiments is that a unit’s potential outcomes depend only on its treatment assignment status. In this session, we analyze cases of interference, or spillovers, where one unit receiving treatment spills over to affect other units. We discuss applications in which spillovers may threaten inference, focusing on quantitative and qualitative methods to uncover and address potential spillovers. These include surveys, interviews, archival research, and focus groups, among others.

Required readings:

Suggested readings:

1:30pm - 3:00pm – Nuts and Bolts of Implementing Natural Experiments: Bundled and Partial Treatments

Even when the assumptions of a natural experiment are met (random assignment to treatment and no spillovers), further challenges remain. Because the researcher does not directly manipulate assignment to treatment—as she would in a true experiment—what precisely the treatment is may be subject to debate and interpretation. In this module, we examine examples of bundled and partial treatments and how researchers may use quantitative and qualitative methods to identify the components of the treatments under analysis.

Required readings:

Suggested readings:

3:30pm - 5:00pm – Designing Natural Experiments: Applications and Placebo Tests

We offer a review of the core approaches to natural experiments, working through a guided review of a problem set (to be completed before class). We then discuss how students might identify and evaluate natural experiments in their own research using a multi-method approach, and how they can apply the same methods to evaluate research conducted by others.

In lieu of required readings, students should complete the assigned problem set for this session.

Suggested readings: