Political Methodology

Political methodology is the branch of political science concerned with how the discipline produces and justifies knowledge.

13 frameworks, from 1880 to today · Political Science · Social Sciences

still a working view ended · each leads to its entry below

Overview

Political methodology is the branch of political science concerned with how the discipline produces and justifies knowledge1. Its practitioners ask not what explains a particular election or regime, but how anyone could know that an explanation is right: what counts as evidence, how concepts should be defined and measured, which research designs can support causal claims, and how qualitative and quantitative evidence can be combined. The stakes are practical. A study claiming that a particular electoral rule causes higher turnout, or that a truth commission reduces violence, is only as good as the methods behind it. Political methodology is the field that builds, tests, and argues about those methods.

The Institutional and Legal Roots of the Field

Political science grew out of law, history, and philosophy, and its earliest practitioners studied institutions by describing them. In the late nineteenth and early twentieth centuries, scholars such as Woodrow Wilson and James Bryce wrote about constitutions, legislatures, and parties by tracing their formal rules and comparing their legal structures. This approach, later called old institutionalism, treated institutions as rules written down and asked how they worked in practice2. Its method was largely descriptive and historical: read the statutes, trace the debates, compare the texts. It produced rich accounts of how governments were formally organized, but it had little to say about whether those arrangements caused the outcomes observers attributed to them, and it offered no systematic way to test claims against evidence beyond the case at hand.

Behavioralism and the Turn to Systematic Evidence

After the Second World War, a generation of scholars pushed political science toward the methods of the natural sciences3. Behavioralism held that political inquiry should focus on observable behavior—voting, protesting, joining—and should test hypotheses with systematic data rather than legal description4. The behavioral revolution drew on survey research, which had been developed to study public opinion during the war, and on statistics. The American National Election Studies, begun in 1948, gave scholars a recurring, comparable body of survey data on voters, and the field's center of gravity shifted toward measurement and inference.

Behavioralism's central methodological claim was that political regularities could be discovered and stated as general laws, much as in physics. Its critics, including some behavioralists themselves, pointed out that the approach struggled with concepts: what exactly is "power," and how would one measure it? This question of concept formation and measurement became a methodological problem in its own right. Giovanni Sartori's work in the 1970s on how to define and compare concepts across cases—warning against "conceptual stretching," in which a term like "democracy" is widened until it fits everything and explains nothing—gave the field a vocabulary for thinking about what its variables actually were.

The Comparative Method and the Problem of Few Cases

Comparative politics faced a distinctive methodological problem: it often had many questions and few cases. A scholar comparing the collapse of authoritarian regimes across a dozen countries could not run a controlled experiment, and a dozen cases were too few for the statistical techniques of survey research. The comparative method that developed in response addressed this by treating cases as configurations to be compared systematically. Arend Lijphart's 1971 essay on the comparative method named the problem and defended the careful comparison of a small number of cases as a legitimate form of inference, distinct from both the single case study and the large-N statistical study.

The most influential solution to the small-N problem was qualitative comparative analysis (QCA), introduced by Charles Ragin in the late 1980s. QCA treats cases as combinations of conditions and uses Boolean logic to identify which combinations are sufficient or necessary for an outcome. Its appeal was that it formalized the comparative intuition that outcomes often depend on several conditions occurring together—a configuration—rather than on any single variable acting alone. The approach, sometimes called case-oriented configurational analysis, gave small-N researchers a transparent procedure, though it also drew criticism for how sensitive its results could be to coding decisions.

The Potential Outcomes Framework and the Credibility Revolution

The most consequential methodological development of the late twentieth century was the importation of the potential outcomes framework from statistics into political science. The framework, associated with the statistician Donald Rubin and building on earlier work by Jerzy Neyman, defines a causal effect as the difference between what happens to a unit under treatment and what would have happened to the same unit under control5. Since no unit can be observed in both states, the framework's central problem is finding a credible comparison group—a "counterfactual" that stands in for the unobserved outcome.

This framework reframed old questions. It made clear that the credibility of a causal claim depends on the research design, not on the sophistication of the statistical model applied afterward. The result was what the field came to call the "credibility revolution": a shift toward designs that mimic experiments. Experimental and quasi-experimental design moved to the center of the discipline. Field experiments, in which researchers randomly assign a treatment—such as a get-out-the-vote mailer or a candidate's campaign message—became a standard tool, and quasi-experimental methods such as regression discontinuity, difference-in-differences, and instrumental variables were used to exploit natural sources of variation when randomization was impossible.

The framework also clarified what was wrong with older practices. A regression controlling for many variables could not, by itself, establish that a relationship was causal, because the controls might not capture the relevant differences between treated and untreated units. The potential outcomes approach made this a matter of design: the question became whether the comparison group was genuinely comparable, and the answer had to be argued rather than assumed.

The Unified Logic of Inference and the Qualitative-Quantitative Divide

In 1994, Gary King, Robert Keohane, and Sidney Verba published Designing Social Inquiry, which argued that qualitative and quantitative research share a unified logic of inference. Both, they claimed, aim at descriptive and causal inference, and both face the same problems of bias, efficiency, and uncertainty. The book's provocation was that qualitative researchers should be held to the same standards as quantitative ones—that a case study, like a regression, must confront the problem of how many observations it has and whether its conclusions are warranted.

The book set off a long and productive argument. Qualitative researchers responded that their work had its own logic, centered on mechanisms and case-based understanding rather than on estimating average effects across populations. This debate produced a wave of methodological work on how qualitative evidence actually supports causal claims. Process tracing, systematized in the 2000s and 2010s, became the leading answer: the close examination of within-case evidence—documents, interviews, sequences of events—to test whether a hypothesized causal mechanism was actually at work. Its advocates distinguished different kinds of tests, such as "hoop tests" and "smoking gun" tests, by how much passing or failing them would tell us about a hypothesis6.

Interpretive and Ethnographic Alternatives

Not all political methodologists accepted the causal-inference frame. Interpretive and ethnographic methodology held that political life is constituted by meanings, and that understanding politics requires grasping how actors understand their own world rather than measuring variables across cases7. This tradition, drawing on philosophy and anthropology, insisted that the choice of methods is itself a claim about what politics is7. Its practitioners developed their own standards of rigor—reflexivity, thick description, attention to the researcher's position—and argued that the unified-logic approach flattened important differences between explaining and understanding.

Multi-Method Research and the Present

By the 2010s, the field had settled into a productive pluralism. Multi-method research—combining large-N statistical or experimental work with case studies or process tracing—became a recognized ideal, on the argument that different methods have different strengths and their combination can compensate for each one's weaknesses. A statistical finding that a program had an effect could be paired with process tracing to show how the effect came about; a case study could generate a hypothesis that a large-N design then tested.

The most recent turn is computational text and network analysis. The digitization of legislative records, social media, and historical archives, combined with machine learning and natural language processing, has given political methodologists new ways to measure concepts and map relationships at scale. Text analysis can classify millions of documents; network analysis can trace how information and influence flow through institutions and online communities. These methods raise their own questions—about measurement validity, about the opacity of machine-learning models, and about whether the patterns they find are substantively meaningful—that are recognizably the field's oldest questions in new form.

Today political methodology is a field of competing and complementary approaches: the potential outcomes framework and experimental design remain dominant for causal questions; process tracing and comparative methods remain central for case-based work; interpretive and ethnographic approaches continue to argue for a different conception of inquiry; and computational methods are expanding what can be measured. The open questions are the ones the field has carried from the beginning: how to define and measure political concepts, how to make credible causal claims when experiments are impossible, and how to combine the precision of large-N analysis with the depth of case-based understanding.

References

  1. Political methodology, Wikipedia
  2. New institutionalism (History), Wikipedia
  3. Political science, Wikipedia
  4. Behavioralism, Wikipedia
  5. Rubin causal model, Wikipedia
  6. Process tracing, Wikipedia
  7. Antipositivism, Wikipedia

The frameworks of Political Methodology

Each framework is a way of working the field has used: when it held, and the question it answers.

  1. 1880-1950

    Old Institutionalism

    How do formal institutions work in practice?

    Fly to its tower in Nousfarer

  2. 1945-1975

    Behavioralism

    What regularities govern observable political behavior?

    Fly to its tower in Nousfarer

  3. 1957-Present

    Formal Theory and Game-Theoretic Modeling

    What follows from rational actors' strategic choices?

    Fly to its tower in Nousfarer

  4. 1970-Present

    Concept Formation and Measurement

    How should political concepts be defined and measured?

    Fly to its tower in Nousfarer

  5. 1970-Present

    Interpretive and Ethnographic Methodology

    How do actors understand their own political world?

    Fly to its tower in Nousfarer

  6. 1971-Present

    Comparative Method

    How can a few cases support causal inference?

    Fly to its tower in Nousfarer

  7. 1987-Present
  8. 1990-Present

    Rubin causal model

    What is a causal effect and how is it identified?

    Fly to its tower in Nousfarer

  9. 1994-Present

    Unified Logic of Inference

    Do qualitative and quantitative research share one logic?

    Fly to its tower in Nousfarer

  10. 1995-Present
  11. 2000-Present

    Process Tracing

    How does within-case evidence test causal mechanisms?

    Fly to its tower in Nousfarer

  12. 2005-Present

    Multimethodology

    How can different methods be combined productively?

    Fly to its tower in Nousfarer

  13. 2010-Present

    Computational Text and Network Analysis

    What can large text and network data reveal about politics?

    Fly to its tower in Nousfarer

How the ideas connect

Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.

Old InstitutionalismBehavioralismFormal Theory and Game-Theoretic Mode…Concept Formation and MeasurementInterpretive and Ethnographic Methodo…Comparative MethodQualitative Comparative AnalysisRubin causal modelUnified Logic of InferenceExperimental and Quasi-Experimental D…Process TracingMultimethodologyComputational Text and Network Analys…

Beyond the field

Where Political Methodology's ideas came from in other fields, and what they led to there: the strongest 12 of 54 links.

A city in Nousfarer: a red-carpeted stair climbing terraces to a crowned tower on an acropolis, more towers round it, forest and snowy mountains beyond.
A city in Nousfarer, the game played in the browser.

Political Methodology is a planet in Nousfarer

Its city has a tower for each of these 13 frameworks. Land, walk in, and climb one concept at a time.

Land on Political Methodology

Related fields

Fields its ideas are linked to

More in Political Science

See where Political Methodology sits on the Map of Human Knowledge