IQMR 2026

Multi-Method Research (Modules 26, 30, 34)

Wednesday, June 24; Thursday, June 25; Friday, June 26

Eggers Hall, Room 032

Jaye Seawright (Northwestern University)

This module sequence explores the productive integration of qualitative and quantitative methods, encompassing case studies, statistical modeling, experiments, and machine learning.

Module 1: Foundations of Multi-Method Design The first session establishes the groundwork for mixed-method research, focusing on designs that pair case studies with regression-based statistical techniques. We will analyze the various roles qualitative components play in strengthening quantitative findings, specifically regarding research designs for testing assumptions related to measurement, confounding, and hypothesized causal paths. Additionally, we will evaluate best practices for selecting cases from a larger population to ensure inferential validity.

Module 2: Experiments and Natural Experiments The second session examines multi-method research within the context of random (or “as-if” random) assignment. We will discuss how to effectively design and pair case studies with experimental and natural-experimental frameworks. This session further refines case selection strategies, adapted specifically for the constraints and opportunities provided by experimental data.

Module 3: Machine Learning and Advanced Integration The final session introduces machine learning tools as both a centerpiece of mixed-method designs and as a support for case studies. We will explore how statistics and machine learning can bolster causal inferences derived from process tracing, as well as their utility in non-causal research goals, such as concept formation and measurement.

Each module includes hands-on activities (ranging from in-room focus groups to guided exercises in R) and practical applications to bridge theory and practice.

Book to Purchase: Seawright, Jason. 2016. Multi-Method Social Science: Combining Qualitative and Quantitative Tools. Cambridge: Cambridge University Press.

Participants may enter the module sequence after it has begun, but their doing so is discouraged.

Foundations of Multi-Method Design (M26, June 24)

This module works through multiple ideas about how to combine qualitative and quantitative research techniques within a single project, working through these concepts with an eye to applications that use regression and similar techniques (e.g., logit, probit, multilevel models) as the quantitative side of an overall design. The goal is to explore optimal research design choices, consider potential weaknesses of such designs, and encounter ideas at the cutting edge of methodological thought in the relevant research traditions.

8:45am - 10:15am – Why Integrate? Roles of Qualitative Work

This session introduces major paradigms of mixed- and multi-method research, including iteration, triangulation, integration, and more. We will discuss the foundational beliefs of each paradigm regarding qualitative and quantitative research and their interrelation, as well as the pragmatic implications of each approach for combining methods.

Required readings:

Suggested readings:

1:30pm - 3:00pm – Building Cases into Regression Designs

This session discusses what is known about the strengths and weaknesses of regression-type research and process-tracing qualitative case studies for causal inference. It then explores specific research design strategies for combining these methods in ways that minimize these weaknesses while enhancing the strengths of each method. We will also discuss case selection.

Required readings:

Suggested readings:

3:30pm - 5:00pm – Hands-On Lab: Regression, Case Studies, and Case Selection

This session is an applied lab in which individual and group activities create a space to experiment with and explore the advantages and limitations of the methodological ideas discussed in today’s sessions while they are fresh. There will be a mix of statistical activities presented in R, qualitative tasks using online primary- and secondary-source research, and group discussion.

Experiments and Natural Experiments (M30, June 25)

This module extends the ideas about mixed- and multi-method design to contexts beyond regression, including natural experiments and laboratory/survey/field experiments.

8:45am - 10:15am – Natural Experiments and Case Studies

This session explores multi-method designs built around natural experiments, considering classic natural experiments, instrumental-variables designs, regression-discontinuity designs, and more. For each, we will ask what central assumptions enable causal inference, whether and how qualitative evidence can enrich tests of those assumptions, and how to design effective qualitative components to provide the most relevant evidence. We will consider famous examples of each type of design, asking if qualitative evidence does support it, or (if not) could have tested its assumptions. We will also discuss case selection.

Required readings:

Suggested readings:

1:30pm - 3:00pm – True Experiments and Multi-Method Integration

This session asks how multi-method design can work with research where the quantitative component involves experimental research. Such projects are an increasingly important part of social science, and the design implications are different in interesting ways from those raised by regression. This session explores designs that engage with those differences, including designs focused around ideas of experimental realism, network and equilibrium effects, and selecting/designing a treatment.

Required readings:

Suggested readings:

3:30pm - 5:00pm – Hands-On Lab: Adding Qualitative Work to Natural/True Experiments

This session is an applied lab in which individual and group activities create a space to experiment with and explore the advantages and limitations of the methodological ideas discussed in today’s sessions while they are fresh. We do not have the time, human-subjects approval, or resources to implement full-scale experiments live online or in the world this afternoon, but we can and will design multi-method experiments, as well as looking at how to add case-study tests to published natural experiments. Once again, there will be a mix of statistical activities presented in R, qualitative tasks including interview/focus group design, and group discussion.

Machine Learning and Advanced Integration (M34, June 26)

Today’s sessions push toward the cutting edge of mixed- and multi-method research design, incorporating machine learning and AI into our thinking about process tracing, case interpretation, and qualitative research, and discussing ways that mixes of these methods and other statistical techniques with qualitative knowledge can help meet longstanding qualitative goals regarding concept formation and measurement.

8:45am - 10:15am – Machine Learning as Support for Process Tracing and Interpretation

Can machine learning and AI supplement established tools of qualitative research without stealing the spotlight? This session suggests two frameworks for such methodological collaboration. First, machine learning methods can search large data sets to find initial clues that become puzzles kicking off qualitative investigations. Second, text-as-data analysis of text collections can highlight especially representative and/or unusual documents, signaling especially productive areas for deep analysis and helping scholars systematically position individual texts within the broader collection for readers and audiences.

Required readings:

Suggested readings:

10:45am - 12:15pm – Multi-Method Designs for Concepts, Measurement, and Theory-Building

This session explores the long-standing, parallel qualitative, quantitative, and statistical/machine learning literatures on description, measurement, concept formation, and theory-building, and asks whether and how these traditions can be mixed in practice to produce better description, measurements, concepts, and theories. Can this earliest stage of research benefit from the same multi-method paradigms that we earlier applied to causal inference?

Required readings:

Suggested readings:

1:30pm - 3:00pm – Hands-On Lab: Machine Learning and Qualitative Evidence for Discovery

This session is an applied lab in which individual and group activities create a space to experiment with and explore the advantages and limitations of the methodological ideas combining machine learning, qualitative methods, and ideas about concept formation and theory-building while they are fresh. We will carry out some small-scale machine learning and text-as-data exercises (reduced in scope to make sure they can execute in our time together) and then use the results to generate qualitative investigations. Group discussion will reveal the ways in which our discoveries are convergent or serendipitously distinctive.