How to Study Political Science: Claims and Cases

To study political science, treat every institution and concept as a claim about what a rule does: who it advantages, what it makes harder, and what changes when the rule changes. Then give most of your hours to testing those claims against real cases. The published mark weightings put analysis and evaluation well ahead of recall, and most revision timetables aim squarely at recall.

The ordering is easy to miss while a course is still running, because what the subject visibly demands is coverage. Constitutions, party systems, treaties, thinkers, election results: the reading list has an end, and getting through it feels like progress. Then a paper comes back carrying a comment about how much it explains and how little it argues, and the graded skill turns out to have been sitting somewhere the reading never touched. Everything below aims at that skill, and most of it is cheap enough to run weekly across a term.

What does a political science exam actually reward?

Work backwards from the weightings, because they are published. AQA lists them for A-level Politics under scheme of assessment, and adds that Ofqual sets the objectives, so every board carries the same three. The numbers therefore describe the subject rather than one specification.

AO1 asks students to demonstrate knowledge and understanding of political institutions, processes, concepts, theories and issues, and carries 30 to 40 percent. AO2 asks them to analyse aspects of politics and political information, including in relation to parallels, connections, similarities and differences, and carries 35 to 45 percent. AO3 asks them to evaluate aspects of politics and political information, including to construct arguments, make substantiated judgements and draw conclusions, and carries 25 to 35 percent.

Read the ranges rather than the labels. Where knowledge is weighted most generously it still stops at 40 percent, and at the other end of the band it is 30, which leaves analysis and evaluation between 60 and 70 percent of the qualification between them. Each of those two names something you do to material rather than something you hold. A plan built around covering the specification is therefore aimed at the one objective that can never exceed 40 percent.

American courses cut the discipline differently and arrive somewhere similar. If the paper closing out your year is a College Board one, its required documents, required cases, and four free-response types run by rules of their own, and scoring a 5 on AP US Government works through them. What follows sits underneath any of those papers.

The discipline is several fields sharing one name

Most departments split political science into roughly four fields, and the split matters because each one wants a different kind of answer.

• Domestic politics. The unit is an institution and the process running through it. Answers turn on how a rule shapes behaviour inside one system.
• Comparative politics. The unit is a case, usually a country, and the answer is almost always a comparison rather than an account of one place.
• International relations. The unit is a state acting among other states, and answers usually rest on a theory of why states behave as they do, applied to an episode.
• Political theory. The unit is an argument, and the question is whether it holds together.

Political theory is where transferred habits fail hardest, and the UNC Writing Center explains why in its handout on writing in political science. It says the field deals primarily with historical and normative, rather than empirical, analysis, and asks writers to attend to two things: the logical consistency of your ideas and the manner in which you use the arguments of other theorists to support your own. That is a different job from assembling evidence, and it sits much closer to what a philosophy paper wants. The reconstruction habits in studying philosophy as arguments and objections move across to a theory module more or less unchanged.

So check which field a given week belongs to before deciding how to revise it. The quietest failure in this subject is a comparative politics module studied as a stack of country profiles, which is a domestic-politics method applied where it does not fit.

Every institution is a rule with a consequence

An institution learned as a description sits inert on the page. The same institution learned as a rule that produces a consequence hands you something to argue with, and arguing is what the marks are for.

Take electoral systems. A description says that first past the post awards a seat to whoever collects the most votes in a constituency. A claim says that this rule pushes the number of viable parties toward two, because a vote for a third party in a district it cannot win buys nothing, and voters and donors both learn that. The claim is Duverger's law, and putting it in that form does three jobs at once: it names the rule, it gives the mechanism, and it makes a prediction somebody could check.

Then check it, because the checking is the part the exam pays for. The United Kingdom runs first past the post and has sent members of well more than two parties to the Commons for decades, and India sustains large multiparty competition under the same rule. Those exceptions are where the evaluation marks live. The mechanism is real inside a single district, and something else, usually the geographic concentration of a party's support, keeps additional parties alive nationally. A student who can say that has an argument. A student who can define first past the post has a definition.

Every large topic in the subject has this shape. More veto points in a system slow policy change. Federal structures produce policy variation between subunits. Proportional systems raise the frequency of coalition government and the bargaining that follows an election. In each case write the rule, the consequence, and a case where it visibly happened, then go looking for the case where it did not. The reasoning runs on the same rails as tracing a model in studying economics through its models, where a shift only means something once you can say what moves next and why.

How do you use countries and elections as evidence instead of describing them?

Comparative politics is where this goes wrong most visibly, because the material arrives pre-packaged as countries and countries invite profiles.

The College Board course in AP Comparative Government and Politics is a clean illustration of the intended use, since it fixes the case set at six: China, Iran, Mexico, Nigeria, Russia, and the United Kingdom. Six countries is few enough that a student can learn each one thoroughly and still answer badly, because the exam is organised by unit rather than by country. Those units are political systems and regimes, political institutions, political culture and participation, party and electoral systems, and political and economic change, and each carries its own published share of the multiple-choice score. The units are the questions. The countries are the evidence.

So build the grid the other way up. Rows are mechanisms: how the executive is chosen and removed, how power is divided between the centre and the regions, how a legislature can and cannot constrain an executive, how a party reaches a ballot, how information moves. Columns are your cases. Revise down a row, and comparison becomes the obvious thing to do, because the row already holds one feature still.

Then keep each comparison controlled. Two cases differing in everything teach nothing, since any of a hundred differences could account for the outcome. A useful pairing holds something steady and varies one feature, so the difference in outcome has somewhere to attach. This is also the boundary with the nearest neighbouring subject: an account of what happened in a country, in order, is a history answer, and the narrative and timeline methods in studying history are built for exactly that. Here the events are evidence for a claim about how systems behave.

Pin down a contested concept before you argue with it

Political science runs on words its own practitioners disagree about. Democracy, power, legitimacy, the state, populism, authoritarian: each carries several defensible definitions, and they do not pick out the same set of real countries. Whether a particular state counts as a democracy depends on which definition you walked in with.

The UNC handout describes the professional response as operationalizing, which it explains as taking important but possibly vague or abstract concepts, such as justice, and defining them in ways that allow them to be measured or tested. Undergraduates are rarely asked to build the measure. Everybody is expected to notice that the choice exists.

The effect on how you study is small and specific. For each contested term keep the two or three main definitions and the case that separates them, since the separating case is the reason the disagreement survives at all. A definition of democracy resting on free and fair elections and a definition that adds effective constraints on whoever wins agree about most countries and disagree about an interesting handful, and exam questions get built out of the handful.

It also changes what an introduction is doing. A definition stated up front commits you, and a marker will hold you to it for the rest of the paper, so pick the one your argument actually needs. Sociology hits the same problem from another direction, where several perspectives explain one social fact and the marks turn on attribution, and keeping each sociological claim with the perspective that owns it deals with that version.

The judgement is the thing being graded

AO3 asks students to evaluate aspects of politics and political information, including to construct arguments, make substantiated judgements and draw conclusions. Substantiated judgements is the operative phrase, and it rules out the most popular essay in the subject: a paragraph for each side, weighed evenly, closing on the observation that there are arguments both ways.

A judgement says which side is stronger and on what grounds. It can be qualified, and usually should be, along the lines of the first account holding under these conditions with this case showing where it stops. The one thing it cannot be is absent.

Style follows from the same standard. Political scientists avoid impressionistic or metaphorical language, or language which appeals primarily to our senses, emotions, or moral beliefs, as the UNC handout puts it, because the persuasion is supposed to arrive through the evidence. In practice that means cutting an adjective and adding a case.

The training for this is short and frequent. One paragraph, most days: a disputable claim, the evidence for it, the strongest objection to it, and a line saying where you land. Fifteen minutes produces something a tutor can respond to, and tutor response is what moves this subject fastest while being the thing students ask for least often. Full timed essays matter closer to the exam, when the skill being drilled is the clock.

Keep a live case file

Political science is one of the few subjects where fresh evidence arrives while you are studying it. An election, a court ruling, a coalition collapsing, a constitutional amendment: each one tests a claim already sitting on your reading list, and each is worth more to you than a textbook example because you watched it happen and can remember the detail.

File by mechanism rather than by date. A ruling goes under how courts constrain executives, alongside the two older cases already there, with one line saying which way it cuts. By the time the paper arrives you are carrying evidence that nobody else in the room has, and an answer built on it reads as thinking rather than as recall of the standard example everyone was handed in week three.

How GeniusPal helps

The four-line note this post keeps pushing is already a question with its answer stapled to it. Given the rule, what consequence follows? Given the consequence, which case shows it happening? Given that case, what does the exception change? Structure like that is why the drilling half of political science converts so readily into practice, and why the mechanism grid is the most valuable file on your machine: rows of mechanisms against columns of cases is a table where every filled cell is a prompt waiting to be asked.

Hand that file to GeniusPal and it writes the questions out of it. A grid exported as a CSV works, and so do a PDF, a Word or PowerPoint file, plain text, and Markdown, up to 10 MB. One set comes back, workable in three modes: quiz, flashcards, or written recall, where you commit an answer before anything is revealed. That third mode earns its keep here, because producing a mechanism with nothing in front of you is as close as a drill gets to the short-answer question that opens most papers.

Two boundaries, stated flatly, since a subject about judgement is a bad place to oversell a tool. First, everything it asks comes out of the file you gave it, so notes that stored a definition with no consequence attached will return cards carrying the identical gap; feed it the four-line version instead and the difference shows on the first pass. Second, nothing here can judge a judgement. Whether your verdict on a contested question is defensible, whether your comparison held the right thing steady, whether an evaluation paragraph earned its conclusion: all three need a human reader, and they are where your own desk hours belong. Free accounts get 2 generations for the lifetime of the account, enough to put one module through and find out whether the questions deserve a slot in your week.