# Political Methodology Hub Political methodology is the subfield that asks how political scientists know what they claim to know — the formal models, statistical tools, research designs, and analytical frameworks that turn political questions into tractable ones. It is the infrastructure subfield: the other four branches of political science generate the questions, and methodology builds the instruments for answering them. Whether a comparativist is measuring [[Democratic Backsliding]], a behavioralist is estimating the effect of media exposure on vote choice, or an IR scholar is modeling deterrence between nuclear powers, the tools they use are methodology's product. For this knowledge base, methodology matters in two ways. The first is that the vault holds a substantial trove of methodological content inside the [[Decision Science Hub]] and its sub-hubs — game theory, statistics, cognitive biases, mental models — that was built without a political-science front door. This hub supplies one, connecting the relevant tools to the political questions they illuminate. The second is that methodological literacy is the vaccine against the abuse of data in political argument. Understanding Selection Bias, Spurious Correlation, Statistical Significance, and the difference between Correlation and causation is what separates genuine analysis from the [[Indicator Cherry-Picking]] and [[Motivated Reasoning]] that characterize much of the political information environment. ## Formal theory and social choice Formal theory uses mathematical models to analyze political situations as strategic interactions among rational actors. Its power is precision: by specifying assumptions explicitly and deriving conclusions logically, formal models reveal what follows from what — and, crucially, what does not follow from the premises that political actors claim. The [[Median Voter Theorem]] is the discipline's most famous formal result: in a two-candidate election where voters have single-peaked preferences on a single dimension, both candidates converge on the median voter's position. The theorem explains a great deal about two-party competition — and its failures explain even more. When the single-dimension assumption breaks down (because politics is about identity, not just policy), when information is asymmetric (because voters don't know where candidates actually stand), or when primary elections pull candidates toward their base before general elections pull them back, the theorem's predictions fail. The failures are themselves informative: the American departure from median-voter convergence — asymmetric polarization, with the [[Republican Party]] moving far rightward — is a research puzzle that formal theory helps to frame. [[Maurice Duverger]]'s law — that single-member-district, plurality-rule elections tend to produce two-party systems — links electoral institutions to party-system structure. The Condorcet Jury Theorem establishes the conditions under which majority rule tracks the truth: if each voter is more likely to be right than wrong, and votes independently, then the probability that the majority is correct approaches certainty as the group grows. The result is both an argument for democracy and a warning about its preconditions. When voters are not independent (because of media manipulation or social pressure) or not individually better than chance (because of systematic [[Disinformation]]), the theorem's optimistic conclusion reverses. ## Game theory Game theory — the mathematics of strategic interaction — is the formal backbone of both international relations and domestic political analysis. The [[Game Theory Hub]] carries the full infrastructure; this section maps the political applications. The [[Prisoners Dilemma]] is the workhorse model: two actors would both benefit from cooperation, but each has an individual incentive to defect, so both defect and end up worse off. It models the [[Arms Race]] dynamic in IR (each side arms because the other might), the Collective Action Problems that plague interest-group organization (each individual benefits from free-riding on others' efforts), and the [[Tragedy of the Commons]] in public-goods provision. The [[Nash Equilibrium]] — the strategy profile from which no player can unilaterally deviate and do better — is the fundamental solution concept. It tells you where strategic interactions settle, and often the answer is a stable outcome that nobody actually likes. The [[Arms Race]], the gridlocked legislature, the race to the bottom in tax competition between states — all are Nash equilibria in which every actor is doing the best they can given what everyone else is doing, and yet the collective result is terrible. This is perhaps game theory's deepest political insight: rational individual behavior can produce collectively irrational outcomes. The structure of the game, not the stupidity of the players, is the problem. The Schelling Focal Point addresses coordination: when multiple equilibria exist, how do actors converge on the same one? The answer is often cultural, historical, or cognitive rather than strictly rational — a finding that links formal theory to the constructivist insight that norms and shared meanings shape strategic behavior. Mechanism Design inverts the standard game-theoretic question: instead of asking what outcome a game produces, it asks what game (what set of rules and incentives) would produce a desired outcome. It is the theoretical foundation for institutional design — constitutions, electoral systems, regulatory frameworks — and for understanding why badly designed institutions produce pathological results. Applied to international relations, game theory structures the analysis of deterrence (the threat to retaliate must be credible), alliance formation (collective defense as a public good with free-rider problems), and the security dilemma (one state's defensive measures are another's offensive threat). The [[Arms Race]] page and the [[Nuclear Deterrence Theory]] page carry the applied content. Applied to domestic politics, game theory models legislative bargaining, coalition formation, and the strategic behavior of courts — any situation where the outcome depends not just on what one actor does but on what all actors do simultaneously. ## Quantitative methods Quantitative political methodology applies statistical tools to political data — surveys, voting records, economic indicators, event counts — to test hypotheses and estimate causal effects. The [[Decision Science Hub]]'s statistics catalog provides the technical foundation: Hypothesis Testing, Statistical Significance, Linear Regression, Regression Analysis, Sample Size, Bayesian Updating, and the Bayesian vs Frequentist divide. The central methodological challenge in political science is causal inference — distinguishing causes from correlations in a world where you cannot run controlled experiments on wars, elections, or constitutional designs. [[Correlation is Not Causation]] is the first principle: just because two things move together does not mean one causes the other. Spurious Correlation is the most common trap: a third variable drives both observed variables, creating the illusion of a causal relationship that does not exist. Selection Bias — the distortion that arises when the cases you study are not representative of the cases you want to generalize about — is the methodological minefield that threatens every comparative or historical study. If you study only democracies that failed, you will overestimate the factors that lead to failure, because you are ignoring all the democracies that had the same factors and survived. The modern causal-inference revolution in political science has developed a toolkit for attacking these problems without laboratory control: natural experiments (situations where nature or accident creates treatment and control groups), instrumental variables, regression discontinuity designs, difference-in-differences, and matching methods. [[Counterfactuals]] — asking what would have happened in the absence of the cause — are the conceptual foundation: a causal claim is always implicitly a comparison between what happened and what would have happened otherwise. When political scientists say that gerrymandering caused a party to win more seats, they mean that in a counterfactual world with fairly drawn districts, the seat distribution would have been different. Making that counterfactual comparison rigorous — rather than merely plausible — is what the causal-inference toolkit is designed to do. ## Survey research and polling Polling — the survey-based measurement of [[Public Opinion]] — is the methodology most visible to the public and the one most subject to political manipulation. The methodological challenges are substantial. Sampling: who gets asked. Question wording: how they get asked. Nonresponse bias: who declines to answer. Social desirability bias: who lies about what they think. And the translation of survey responses into election predictions: likely voter models, aggregation, uncertainty quantification. The growing difficulty of reaching representative samples by phone, combined with the rise of online panels and the political weaponization of poll results, makes survey methodology one of the most active and consequential research frontiers in the discipline. Survey research connects directly to the [[Political Behavior Hub]]: the empirical findings about [[Polarization]], [[Affective Polarization]], [[Political Trust]], and [[Partisan Sorting]] are all based on survey data, and the validity of those findings depends on the quality of the survey methodology that produced them. Sample Size, Selection Bias, and Statistical Significance are the concepts that determine whether a poll is telling you something real about the electorate or merely reinforcing existing narratives. ## Qualitative and comparative methods Not all political questions are best answered with numbers. Qualitative methods — case studies, process tracing, comparative-historical analysis, elite interviewing, archival research — provide the depth, context, and causal-mechanism identification that statistical analysis cannot. When a comparativist asks why democracy failed in the [[Weimar Republic]] or why it survived in India, the answer requires tracing a specific sequence of events, decisions, and structural conditions, not estimating an average treatment effect across a dataset. Process tracing — following the causal chain from putative cause to observed effect through the intervening steps — is the qualitative equivalent of a controlled experiment: it tests whether the hypothesized mechanism actually operated in the case at hand. Did the think-tank memo actually reach the senator's desk before the vote? Did the media coverage actually shift public opinion before the policy change? Process tracing requires showing the links, not just the endpoints. Comparative-historical analysis — the structured comparison of a small number of cases selected for their theoretical relevance — is the method that [[Barrington Moore Jr.]], [[Charles Tilly]], and [[Theda Skocpol]] used to build the foundational arguments in comparative politics. The logic is Mill's methods: find cases that differ on the outcome and ask what else differs, or find cases that share the outcome and ask what else they share. The methodological debate between quantitative and qualitative approaches has largely given way to a mixed-methods consensus: the best political science uses each method where it is strongest — statistics for generalization and hypothesis testing, qualitative methods for mechanism identification and contextual depth — and combines them when possible. This convergence matters for non-specialists because it means that the strongest political-science findings are typically supported by both statistical patterns and detailed case evidence. When the backsliding literature finds statistical associations between executive aggrandizement and democratic erosion *and* traces the causal mechanisms in individual country cases, the combined evidence is substantially more persuasive than either type alone. ## Data science and political analytics The explosion of available data — social media posts, satellite imagery, government records, text corpora, geospatial data — has created a computational turn in political methodology. Text analysis (automated coding of speeches, legislation, news coverage), network analysis (mapping connections among politicians, donors, and organizations), and machine learning (prediction and classification at scale) are increasingly standard tools. For this vault, the data-science dimension connects directly to the empirical content: the 2026 Midterms race files, the [[Authoritarianism and Democratic Backsliding Timeline]], the economic-indicator dashboard documented in the [[Economics Hub]], and the donor-network mappings are all exercises in applied political data analysis. The methodological principles — transparency about data sources, honesty about uncertainty, vigilance against the biases that algorithms inherit from their training data — are what distinguish genuine analysis from the datafied [[Propaganda]] that saturates the political information environment. ## Key figures [[Maurice Duverger]] — the institutional analysis of electoral rules and party systems. The social-choice tradition runs from Condorcet and Borda through Kenneth Arrow's impossibility theorem (you cannot design a voting system that satisfies a small set of reasonable conditions simultaneously) to modern Mechanism Design. The game-theoretic tradition runs from von Neumann and Morgenstern through [[Nash Equilibrium]] to the applied models of deterrence, bargaining, and Collective Action Problems that IR and American politics scholars use daily. ## See also [[Political Science Hub]] | [[Political Behavior Hub]] | [[International Relations Hub]] [[Decision Science Hub]] | [[Game Theory Hub]] [[Nash Equilibrium]] | [[Prisoners Dilemma]] | Schelling Focal Point | Mechanism Design [[Median Voter Theorem]] | Condorcet Jury Theorem | Polling Hypothesis Testing | Statistical Significance | Selection Bias