← Introduction to Political Science outline

Week 13 · Chapter

Chapter 13 — Comparative Politics

Introduction to Political Science Free sample edition

The module's primary reading · ~45–55 minutes · Objective 7 (apply the comparative method to political systems and political development — most-similar and most-different designs, state capacity, and the competing explanations of democratization — and read comparative-governance indices critically, with their edition and year), Objective 4 (compare regime types and constitutional structures — direct and representative, electoral and liberal democracy; authoritarian and totalitarian rule; democratization and backsliding; what constitutions do; the rule of law, separation of powers, and checks and balances) · Read it first this week; the lecture, the slides, the tutorial, the quiz, and the workshop all build on it.

In plain language

This week asks why some countries are rich democracies and others are not, and how you would test an answer. You cannot run an experiment on a country, so political scientists compare cases (the countries, regions, or elections they study side by side) chosen on a plan. You will meet four rival explanations for democratization (the process by which a country becomes democratic), and each one has strong defenders and strong critics. You will also read two published governance indices (scores that combine many expert judgments about a country into one number or category) for what they measure and what they do not. You will not be told which explanation wins; you will be shown how to weigh them and how you could test one.

Why you already care about this

Picture two community gardens in neighboring towns. Every detail here is illustrative. They have the same size, 40 plots each, the same soil, the same budget, and 25 volunteers each. After 3 seasons, one garden is thriving and the other has gone to weeds. A neighbor asks you why.

You cannot rerun the three seasons with one thing changed, the way a science class reruns a test. So you do the next best thing. You list what the two gardens share, and you hunt for the one thing that differs. It turns out the thriving garden keeps a shared watering rota, and the other leaves watering to whoever shows up. You have not proved that the rota saved the garden. You have found the best place to look next.

That is the whole method of this chapter, scaled up from two gardens to the countries of the world. Political scientists cannot assign democracy to one country and hold it back from another. They compare cases with care, they argue about what the comparisons show, and they read published ratings such as Freedom House's Freedom in the World 2026 and V-Dem's Democracy Report 2026. This chapter teaches you to do all three.

By the end of this chapter you will be able to:

tell a most-similar design from a most-different design and say what each can and cannot show;
explain the many-variables, small-N problem and name three partial fixes;
place a country on two separate dimensions, state capacity and regime type;
state the four rival explanations of democratization with a proponent's and a critic's case for each;
read a Freedom in the World score and a V-Dem regime count with their edition and year, and say what they do not show.

Carry this line with you: a comparison narrows the suspects — it never signs the confession.

Where we are

Already covered: Week 1 named comparative politics as one of the five subfields and the comparative method as one of your tools. Week 5 sorted regimes and previewed Freedom House and V-Dem. Week 8 gave you the three-layer read. Week 11 read a real election, and Week 12 read a real poll and kept correlation apart from cause.
This week: the comparative method's two designs, the many-variables, small-N problem, state capacity against regime type, four explanations of democratization, and two governance indices read in full.
How they connect: Week 12's habit of asking "correlation or cause?" is the habit this week turns on the biggest question in comparative politics.
Where we're going next: Week 14 turns from politics inside countries to politics between them, with international relations.

Three earlier ideas carry this chapter. From Week 5, an electoral democracy has free, fair, competitive elections that incumbents can lose, and a liberal democracy adds protected rights, a free press, independent courts, and the rule of law. From Week 8, the design layer is what the rules say, the practice layer is what actually happens, and the verdict layer is what ought to happen. From Week 12, a correlation is an association between two things, which by itself does not show that one causes the other. Keep all three within reach.

1. Compare on a plan: most-similar and most-different designs

Start with a lab. A chemist who wants to know whether heat speeds a reaction runs two beakers. Everything is the same except the heat. If the warm beaker reacts faster, the heat is the best suspect. That is an experiment: you change one thing and hold everything else fixed.

No one can do that with countries. You cannot give one country a new constitution and keep an identical twin without one. So political scientists choose their cases on a plan instead. A case is one unit a political scientist compares, such as a country, a region, or an election, chosen on a plan rather than at random. Two plans are the field's standard tools, both from Przeworski and Teune (1970), The Logic of Comparative Social Inquiry, taught here in words.

A most-similar systems design is a comparison of cases that are alike in most respects but differ in the outcome, so the one difference that tracks the outcome becomes the explanation to test.
A most-different systems design is a comparison of cases that are unlike in most respects but share an outcome, so the one trait they also share becomes the explanation to test.

The garden story was a most-similar design. The two gardens matched on size, soil, budget, and volunteers, and they differed on one thing, the watering rota.

Worked example — a most-similar design. Doranth and Kesslin are made-up neighbors, and everything about them is illustrative. They match on 4 traits: region, income band, main export, and population size. They differ on 1: Doranth has an independent election commission and Kesslin does not. The outcome differs too, since Doranth has held 3 peaceful handovers of power and Kesslin has held 0. You compared 4 + 1 = 5 traits. The commission becomes the explanation to test next, not a proven cause.

Worked example — a most-different design. Amberlay and Tsendi are made-up countries that differ on 4 traits: region, income band, main export, and legal tradition. They share 1 trait, courts whose judges cannot be removed for their rulings, and they share the outcome, decades of elections the incumbents sometimes lose. You compared 4 + 1 = 5 traits, and the shared trait becomes the explanation to test next.

Notice the phrase both examples end on: the explanation to test next. Each design narrows the list of suspects. Neither one proves anything by itself. The difference you found may be a coincidence, and a factor you never measured may be doing the work. Maybe Doranth's commission and its peaceful handovers both came from something earlier, such as a peace deal you never listed. These designs are tools for narrowing explanations, never proof machines.

Misconception check. Students often think the comparative method means comparing any two countries you like. That is tempting, because "compare" sounds like something anyone can do with any pair. But a comparison without a plan cannot point at anything. If two countries differ in a dozen ways and in their outcome, you have no way to tell which difference matters. The method is the plan: alike in most respects with different outcomes, or unlike in most respects with a shared outcome.

Memory hook: similar cases, find the difference; different cases, find the common thread.

2. Face the core problem: many variables, small N

Imagine you want to know why one of two cafés on your street is always busier. You list the possible reasons: price, location, music, the seating, the wireless internet, the opening hours, the friendliness of the staff. That is seven possible reasons and only two cafés. Every reason you list could explain the difference, and two cafés cannot tell you which one does.

Comparative politics has this problem on a world scale. Lijphart (1971), "Comparative Politics and the Comparative Method", published in the American Political Science Review and taught here in words, gave it a name.

A variable is a feature that differs from one case to another, such as a country's wealth or its electoral rule.
Many variables, small N is the core problem of comparative politics: there are many possible explanations and only a few cases to test them on, so a comparison can narrow the explanations but rarely settle them.

In this course, N is the number of cases in a comparison. Doranth and Kesslin give N = 2.

Lijphart offered partial fixes. First, extend the study across time, so each country supplies several observations instead of one. Second, narrow the property space, the set of variables a comparison tracks, by combining related variables. Third, pick comparable cases, which is the most-similar logic again.

Worked example — extending across time. Doranth and Kesslin give N = 2 cases. Watching each at 3 elections gives 2 × 3 = 6 observations, which is extending across time. The list of possible explanations stays long, so the fix helps without settling the question.

A limitation worth saying out loud: none of the three fixes makes the problem vanish. Six observations from two countries are not six independent countries. They share history, culture, and everything else you did not list.

Misconception check. Students often think that adding more years of data from the same two countries solves the small-N problem. That is tempting, because the count of observations goes up. But each new year from Doranth still carries everything else about Doranth with it. The fix helps you see change within a country; it does not give you a new country to compare.

Memory hook: more years help; they are not more countries.

3. Separate two axes: state capacity and regime type

Think about two delivery companies. The first reaches every street in the region, on time, every day. The second has whole neighborhoods it never reaches. Now ask a different question about each: who picks the company's leaders? In one, the workers elect the board. In the other, one owner decides everything. Reach and leadership are two different questions. A company can score high on one and low on the other.

Countries work the same way.

State capacity is whether a state can tax, deliver services, and enforce its laws across its whole territory, a dimension separate from who rules and how.

Regime type is the other dimension: who rules and how, the question you met in Week 5 with democracy and autocracy. The two are separate, and all four combinations are real. A democracy can be capable or weak. An autocracy can be capable or weak.

Worked example — the capacity grid. Four made-up countries, all illustrative, each placed on both dimensions:

Four made-up countries placed on two dimensions, state capacity and regime type (illustrative)
Made-up country	What the state can do	How leaders are chosen	Cell on the grid
Marrowby	collects taxes across its territory	competitive elections	a capable democracy
Solvane	collects taxes across its territory	no competitive elections	a capable autocracy
Ytrel	cannot run schools outside its capital	competitive elections	a weak democracy
Quoravi	cannot enforce its laws in its border provinces	no competitive elections	a weak autocracy

Two dimensions with two levels each give 2 × 2 = 4 combinations, and each made-up country fills 1 cell. In words: knowing where a country sits on one axis tells you nothing certain about where it sits on the other.

Misconception check. Students often treat state capacity and regime type as one axis, as if a strong state must be a dictatorship and a democracy must be weak. That is tempting, because some famous dictatorships were very good at control. But Solvane and Marrowby both collect their taxes, and only one of them holds competitive elections. A democracy can struggle to collect its taxes, and an autocracy can fail to reach its own border provinces. Ask the two questions separately.

Memory hook: can the state do it, and who decides — two questions, two axes.

4. Test the first explanation: modernization and the correlation trap

Here is a pattern from everyday life. Neighborhoods with more coffee shops tend to have more bike lanes. Do coffee shops cause bike lanes? Probably not. A third thing, a crowded, walkable neighborhood, tends to bring both. Two things that move together are a correlation, and a correlation leaves the cause open.

The oldest explanation of democratization runs into exactly this question.

Democratization is the process by which a country becomes democratic, moving toward free and fair elections and protected rights.
Modernization theory is the explanation that economic development, meaning wealth, education, and urban life, makes democracy more likely to emerge or to last, resting on a robust correlation whose mechanism and direction are still debated.

Lipset (1959), "Some Social Requisites of Democracy", is the classic early statement, taught here in words. Wealth and democracy tend to come together, and that correlation is robust and has been found again and again. That part is an empirical finding, and you can state it plainly.

What is still argued is the arrow. There are three possibilities, and each has defenders.

Wealth might support democracy. Proponents argue that educated, secure citizens demand a say and can organize to get it.
Democracy might support wealth. That would be reverse causation, the possibility that the arrow runs the other way from the one first assumed, as when democracy supports wealth rather than wealth supporting democracy.
A third factor might drive both, the way a walkable neighborhood drives coffee shops and bike lanes.

Misconception check. Students often say that Lipset proved wealth causes democracy. That is tempting, because the correlation is strong and famous. But a correlation is an association, not a cause, and the mechanism and direction are still debated. Lipset documented that wealth and democracy tend to come together. Which way the arrow points, and whether a third factor sits behind both, is the open question.

Memory hook: found together is not proof of cause.

5. Weigh three rivals: institutions, culture, and the resource curse

Modernization is one explanation. Three more compete with it, and this course presents each at full strength with no winner. Each starts from a different place you can picture.

Institutions

Picture two housing co-ops. In the first, any member can propose a rule, every member votes, and the budget is posted for all to read. In the second, the founding family controls the budget and the rules, and other members pay in. Over time the first co-op may attract members and fix its roof, while the second may lose members and let repairs slide. Acemoglu and Robinson (2012), Why Nations Fail, taught here in words, put rules like these first.

Inclusive institutions are political and economic rules that spread power widely, protect rights, and let people use their talents, in the institutional account of why nations prosper.
Extractive institutions are political and economic rules built to transfer resources from the many to a narrow group, in the institutional account of why nations fail.

Proponents argue that institutions explain both wealth and democracy, so the correlation Lipset found is a product of good rules. Critics respond that institutions are hard to measure apart from the outcomes they are meant to explain.

Political culture

Picture two volunteer fire companies. In one, members trust each other and accept a vote they lost. In the other, every lost vote starts a feud. The rules on paper are the same.

Political culture is the widely shared set of beliefs and values about politics in a society, such as trust in others and tolerance of opponents.

Proponents argue that trust and tolerance help a democracy last, because a democracy needs losers who accept the result. Critics respond with the circularity critique, the objection that a cultural explanation of democracy goes in a circle when it measures a democratic culture by democratic behavior and then uses that culture to explain democracy. They add that cultures change faster than the theory expects.

The resource curse

Picture a town council funded entirely by a lottery jackpot. It never has to ask residents for taxes. Does it still have to listen to them as closely?

The resource curse is the claim that wealth from oil and minerals tends to come with less democracy, a pattern whose strength and mechanism scholars still dispute.

It comes from Ross (2001), "Does Oil Hinder Democracy?", taught here in words. Proponents argue that rulers who live on resource income need less consent from taxpayers. Critics respond that the strength of the pattern depends on which countries and years are counted, and on the institutions in place before the discovery.

Misconception check. Students often pick the explanation that sounds most convincing and treat it as the answer. That is tempting, because a good story feels like proof. But each of the four explanations rests on evidence, and each faces a serious critique. Stating a theory's strongest case is not the same as declaring it the winner. In this course the verdict stays with you, and it has to be argued from evidence.

Memory hook: four explanations, four critiques, no winner handed to you.

6. Read Freedom in the World 2026: what an index measures

Think about the score posted in a restaurant window after a health inspection. Trained inspectors walk through with a published checklist: food temperatures, hand-washing stations, storage. They score each item and add them up. The score is useful, and you can check how it was built. It is still a judgment, made by people, against a checklist someone chose. And it tells you the kitchen's condition on inspection day, not why the kitchen is that way.

A governance index works the same way.

A governance index is a score that combines many expert judgments about a country's political conditions into one number or category, published with a stated method.
An indicator is one specific question an index scores for each country, such as whether elections are free, which is combined with the other questions into the total.
Expert coding is trained people scoring each indicator for each country against a published method, which makes an index transparent and still a judgment process.
An edition is the year an index is published, which is written beside the year its scores cover because the two are usually different.

The index you read first is Freedom House's Freedom in the World 2026, titled The Growing Shadow of Autocracy, covering 2025, for 195 countries and 13 territories. You write it as Freedom in the World 2026, covering 2025: the edition and the year it covers, together.

Worked example — the scale. Expert coders score 25 indicators: 10 for political rights and 15 for civil liberties. Each indicator gets 0 to 4. So 10 political-rights indicators × 4 = 40, and 15 civil-liberties indicators × 4 = 60. The total runs 40 + 60 = 100, from 10 + 15 = 25 indicators. In words: political rights can add to 0 to 40, civil liberties to 0 to 60, and the total runs 0 to 100. The total sorts a country as Free, Partly Free, or Not Free, written exactly as the index spells them.

Worked example — the four pinned scores. From Freedom in the World 2026, covering 2025:

Freedom in the World 2026 scores for four countries, covering 2025: political rights (0 to 40), civil liberties (0 to 60), total (0 to 100), and status
Country	Political rights	Civil liberties	Total	Status
Canada	39	58	97	Free
Japan	40	56	96	Free
Botswana	31	44	75	Free
Mexico	26	32	58	Partly Free

Add the parts and you reach each total: Canada 39 + 58 = 97, Free; Japan 40 + 56 = 96, Free; Botswana 31 + 44 = 75, Free; Mexico 26 + 32 = 58, Partly Free.

Then read the parts. Japan's political rights sit at 40 of 40, the top of that range. Canada's sit at 39, which is 40 − 39 = 1 below the top, and its civil liberties sit at 58, which is 60 − 58 = 2 below. Botswana's total sits 75 − 58 = 17 above Mexico's on the 0 to 100 scale.

Worked example — the headline. In this edition, 88 of 195 countries were rated Free, which is 88 ÷ 195 × 100 ≈ 45.1 percent of countries, a share of countries, not of people. That leaves 195 − 88 = 107 countries rated Partly Free or Not Free. Global freedom declined for the twentieth consecutive year. In 2025, 54 countries declined and 35 improved, so 54 − 35 = 19 more countries declined than improved. Bolivia, Fiji, and Malawi moved from Partly Free to Free.

In practice (your tools). In this week's workshop you rebuild the scores in a spreadsheet. Put each country's political-rights score in column B and its civil-liberties score in column C, starting in row 2, and type =B2+C2 in D2 to rebuild the total, then fill it down. To turn the Free count into a share of countries, type =88/195*100 in an empty cell and round the result to one decimal place. Every value you type comes from the Freedom House pages, never from memory.

Now the read, with the data scaffold you have used since Week 11. What is measured: the conditions for political rights and civil liberties, as expert coders scored them. Over whom and when: 195 countries and 13 territories, in 2025. What it shows: a score and a status for each, and a count of how many moved. What it does not show: why any country scored as it did. Correlation or cause: neither, by itself. A score is a reading of conditions.

Misconception check. Students often read an index score as a raw physical fact, like a temperature, and read a low score as a verdict on a country's people. Both are tempting, because a number looks exact and a ranking feels like a grade. But a score is the output of expert coders applying a published method, so it is transparent and still a judgment. It describes conditions in a year and is never a verdict on a people. And you cannot read causes off the numbers: Botswana's 75 does not tell you why it scored 75.

Memory hook: an index reports the kitchen, not the recipe.

7. Set V-Dem beside it: categories follow the source

Go back to the restaurant. The health inspector posts a number. A food critic posts stars. Both are judging the same kitchen, and their scales do not convert into each other. Four stars is not "a 96". You would never average the two.

V-Dem's Democracy Report 2026, titled Unraveling the Democratic Era?, is the second index. It counts regimes for 2025 in four categories, and you write them exactly as the report spells them: 31 liberal democracies, 56 electoral democracies, 57 electoral autocracies, and 35 closed autocracies.

An electoral autocracy is a regime category in which multiparty elections are held but fall short of free and fair, so rulers face a contest they are unlikely to lose.
A closed autocracy is a regime category in which no multiparty elections choose the top offices at all.
Autocratization is movement away from democracy, whatever a country's starting point, measured as a decline in an index's democracy scores.

The report also states that 74 percent of the world's population lives in autocracies and that 44 countries are autocratizing.

Worked example — the V-Dem headline. From the Democracy Report 2026, for 2025: 31 + 56 = 87 democracies and 57 + 35 = 92 autocracies, for 87 + 92 = 179 countries counted. Of the world's population, 74 percent lived in autocracies, which leaves 100 − 74 = 26 percent outside them. Read that carefully: 92 autocracies versus 87 democracies is a close count of countries, yet the share of people is far from even. A count of countries and a share of people are different bases, and a share always names its base.

Worked example — the cross-check. Freedom House rated 88 countries Free, and V-Dem counted 87 democracies. The two numbers sit close, and they are still different measures. They use different categories and count different sets of countries, 195 against 179, a difference of 195 − 179 = 16. You never merge them.

One caveat to carry: two careful indices can disagree about a single country, because they ask different questions and weigh them differently. When they disagree, you report both, each in its own words, with its edition and year.

Misconception check. Students often treat "Free" and "liberal democracy" as two words for one thing, and add or average figures from the two indices. That is tempting, because 88 and 87 sit so close together. But Freedom House's statuses and V-Dem's categories come from different methods, and the two sources do not even count the same set of countries. Categories follow the source: never relabel, merge, or rank one source's categories against another's.

Memory hook: stars are not inspection scores — keep each scale in its own words.

8. Weigh the explanations: which one, and how would you test it?

Back to the two gardens, with their matching soil and their very different seasons. Suppose four neighbors each offer a different reason for the weeds: the watering rota, the town's wealth, the volunteers' habits, and a grant that meant no one had to fundraise. Each neighbor can point to something real. The question is which reason holds up when you test it against other gardens.

That is this week's open question. Which explanation best accounts for why some countries are rich democracies and others are not, and how would you test it? Here is each side at full strength, for you to weigh.

Proponents of modernization point to a correlation between wealth and democracy that has held up across many studies.
Proponents of institutions argue that rules about power come first, and that inclusive rules produce both wealth and democracy.
Proponents of political culture argue that values make rules stick, so the same constitution works in one place and fails in another.
Proponents of the resource curse argue that where a government's money comes from shapes whom it must answer to.

Each side names a real weakness in the others. Modernization's critics ask about the arrow. Institutionalists face the measurement critique. Cultural accounts face the circularity critique. The resource curse faces the question of which countries and years are counted.

Keep the kinds of claim apart as you weigh. "Canada scored 97 in Freedom in the World 2026" is empirical, checkable against the country page. "Wealth tends to come with democracy" is empirical, a correlation. "The index ought to weight civil liberties more heavily than political rights" is normative, a claim about how the measure should be built, and no score settles it. Correlation or cause: the indices report conditions in 2025 and show no cause. The link between wealth and democracy is a correlation whose mechanism and direction are still debated.

Pause and think, choosing one of these two:

(a) Pick the explanation you find least convincing and state its strongest case, as its best defender would state it.
(b) Pick any one explanation and design a test for it: say whether you would use a most-similar or a most-different design, and describe the two cases you would want.

Misconception check. Students often think that a strong index score, or a strong correlation, settles which explanation is right. That is tempting, because numbers feel final. But the indices describe conditions, and the correlation describes an association. Choosing an explanation means arguing from evidence about causes, and that argument is yours to make.

Memory hook: the data narrow the field; your reasons pick the runner.

Key terms

case — One unit a political scientist compares, such as a country, a region, or an election, chosen on a plan rather than at random. (first used: 1. Compare on a plan: most-similar and most-different designs)
most-similar systems design — A comparison of cases that are alike in most respects but differ in the outcome, so the one difference that tracks the outcome becomes the explanation to test. (first used: 1. Compare on a plan: most-similar and most-different designs)
most-different systems design — A comparison of cases that are unlike in most respects but share an outcome, so the one trait they also share becomes the explanation to test. (first used: 1. Compare on a plan: most-similar and most-different designs)
variable — A feature that differs from one case to another, such as a country's wealth or its electoral rule. (first used: 2. Face the core problem: many variables, small N)
many variables, small N — The core problem of comparative politics: there are many possible explanations and only a few cases to test them on, so a comparison can narrow the explanations but rarely settle them. (first used: 2. Face the core problem: many variables, small N)
property space — The set of variables a comparison tracks; narrowing it by combining related variables is one partial fix for having too few cases. (first used: 2. Face the core problem: many variables, small N)
state capacity — Whether a state can tax, deliver services, and enforce its laws across its whole territory, a dimension separate from who rules and how. (first used: 3. Separate two axes: state capacity and regime type)
democratization — The process by which a country becomes democratic, moving toward free and fair elections and protected rights. (first used: 4. Test the first explanation: modernization and the correlation trap)
modernization theory — The explanation that economic development, meaning wealth, education, and urban life, makes democracy more likely to emerge or to last, resting on a robust correlation whose mechanism and direction are still debated. (first used: 4. Test the first explanation: modernization and the correlation trap)
reverse causation — The possibility that the arrow runs the other way from the one first assumed, as when democracy supports wealth rather than wealth supporting democracy. (first used: 4. Test the first explanation: modernization and the correlation trap)
inclusive institutions — Political and economic rules that spread power widely, protect rights, and let people use their talents, in the institutional account of why nations prosper. (first used: 5. Weigh three rivals: institutions, culture, and the resource curse)
extractive institutions — Political and economic rules built to transfer resources from the many to a narrow group, in the institutional account of why nations fail. (first used: 5. Weigh three rivals: institutions, culture, and the resource curse)
political culture — The widely shared set of beliefs and values about politics in a society, such as trust in others and tolerance of opponents. (first used: 5. Weigh three rivals: institutions, culture, and the resource curse)
circularity critique — The objection that a cultural explanation of democracy goes in a circle when it measures a democratic culture by democratic behavior and then uses that culture to explain democracy. (first used: 5. Weigh three rivals: institutions, culture, and the resource curse)
resource curse — The claim that wealth from oil and minerals tends to come with less democracy, a pattern whose strength and mechanism scholars still dispute. (first used: 5. Weigh three rivals: institutions, culture, and the resource curse)
governance index — A score that combines many expert judgments about a country's political conditions into one number or category, published with a stated method. (first used: 6. Read Freedom in the World 2026: what an index measures)
indicator — One specific question an index scores for each country, such as whether elections are free, which is combined with the other questions into the total. (first used: 6. Read Freedom in the World 2026: what an index measures)
expert coding — Trained people scoring each indicator for each country against a published method, which makes an index transparent and still a judgment process. (first used: 6. Read Freedom in the World 2026: what an index measures)
edition — The year an index is published, which is written beside the year its scores cover because the two are usually different. (first used: 6. Read Freedom in the World 2026: what an index measures)
electoral autocracy — A regime category in which multiparty elections are held but fall short of free and fair, so rulers face a contest they are unlikely to lose. (first used: 7. Set V-Dem beside it: categories follow the source)
closed autocracy — A regime category in which no multiparty elections choose the top offices at all. (first used: 7. Set V-Dem beside it: categories follow the source)
autocratization — Movement away from democracy, whatever a country's starting point, measured as a decline in an index's democracy scores. (first used: 7. Set V-Dem beside it: categories follow the source)

Summary

A most-similar design compares like cases with different outcomes; a most-different design compares unlike cases with a shared outcome.
Both designs narrow the suspects; neither proves a cause.
Many variables, small N is the field's core problem; extending across time, narrowing the property space, and picking comparable cases help without solving it.
State capacity and regime type are two separate axes, and all four combinations are real.
Modernization rests on a robust correlation between wealth and democracy; its mechanism and direction are still debated.
Institutions, political culture, and the resource curse compete with modernization, and each faces a serious critique.
Freedom in the World 2026, covering 2025, scores 25 indicators from 0 to 4 for a total of 0 to 100 and rated 88 of 195 countries Free.
V-Dem's Democracy Report 2026 counts 31, 56, 57, and 35 regimes for 2025 and puts 74 percent of the world's population in autocracies.
An index reports conditions in a year; it shows no cause and is never a verdict on a people.
Categories follow the source: never merge Free with liberal democracy.

Check your understanding

Work these before looking at the answers — they are ungraded, and the answers (with reasoning) are at the bottom of the page.

A researcher studies two made-up island provinces that share a language, a main industry, a population of about the same size, and the same national laws. One province has high voter turnout and the other has low turnout, and the only other difference she finds is that the high-turnout province mails ballots to every voter. Which design is this, and what can she conclude?
A made-up city-services index scores 8 indicators, each from 0 to 5. What is the top possible total? If a city scores 26, how far below the top is it? Show the steps.
A made-up country, Halloway, holds competitive elections that incumbents often lose, but its government cannot collect taxes in two of its five regions. Place it on the capacity grid.
Sort each statement as an empirical claim about a correlation, an empirical claim about a cause backed by an argument, or a normative claim: (a) "Across a set of made-up countries, those with more paved roads tend to hold more elections." (b) "Countries ought to be ranked only on civil liberties." (c) "Paved roads cause elections."

Answers (worked reasoning — read after attempting)

It is a most-similar design: the two provinces are alike in most respects and differ in the outcome, turnout. She can conclude that mailed ballots are the explanation worth testing next. She cannot conclude that mailed ballots caused the higher turnout, because a difference she did not measure might be doing the work, and with N = 2 she has many possible explanations and few cases.
The top total is 8 × 5 = 40. A city that scores 26 sits 40 − 26 = 14 below the top. In words: eight questions worth up to five each give a 0 to 40 scale, and 26 is fourteen short of a perfect score.
Halloway is a weak democracy. On regime type, its competitive elections that incumbents lose place it with the democracies. On state capacity, failing to collect taxes in two of its five regions places it on the weak side. The two answers come from two separate questions.
(a) is an empirical claim about a correlation: it says two things tend to come together, and you could check it against the data. (b) is normative: it says how countries ought to be ranked and has to be argued from values. (c) is a causal claim with no argument behind it. The correlation in (a) does not show it, because a third factor, such as overall wealth, could bring both roads and elections. This is the correlation trap from section 4.

Next up in this module: the Readings & Resources page, the lecture tutorial, the practice set, and the quiz. Next week, the realism model and its two rivals take the comparative lens to international relations.

Built with The Course Maker · samples.thecoursemaker.com