Freedom in the World

• 1. Introduction

  ›

  Freedom in the World assigns every country and territory on Earth a score. That score can sink as low as -4 or climb all the way to 100. Freedom House, the U.S.-based non-governmental organization behind the report, has turned that number into one of the most widely cited measures of democracy anywhere. Political researchers use its rankings as a source in their own work. Critical assessments of the survey have also pointed to potential biases stemming from its funding, its methodology, its ideological leanings, and the subjective nature of the scoring. Freedom House has defended its ratings as independent and evidence-based. So how does a country actually earn its number? What separates a country from a territory in the survey's eyes? And how strong is the case its critics have built against it?

• 2. Origins And Categories

  ›

  Raymond Gastil launched Freedom in the World in 1973. Each year since, the survey has given every state and territory two ratings. The scale runs from 1, the freest, to 7, the least free. Freedom House sorts every entry into one of three classifications: Free, Partly Free, or Not Free. Researchers who study democracy broadly rely on the resulting numbers. The survey correlates highly with other measures of democracy, including the Polity data series, a separate index scholars use to track political systems over time. Those correlations only measure how closely two finished numbers agree. Building the number in the first place is a far more granular process.

• 3. Building The Score

  ›

  Each country and territory receives a score of 0 to 4 on 25 separate indicators. Ten of those indicators cover political rights, and 15 cover civil liberties. The civil liberties total can run from 0 to 60. The political rights component ranges from -4 to 40, because a separate discretionary question can subtract additional points. Before the 2018 report, a different discretionary question worked the opposite way. It could add 1 to 4 points to certain traditional monarchies. Freedom House defined these as governments with 'no political parties or significant electoral processes' that still 'provided for some form of consultation with the public.' That kind of assessment is now folded into other parts of the methodology. The raw scores convert into the 1 to 7 ratings, and the average of a country's two ratings sets its overall status. Free covers 1.0 to 2.5, Partly Free spans 3.0 to 5.0, and Not Free runs 5.5 to 7.0. Since 2020, Freedom House has played down the ratings themselves in its published reports, even though the underlying numbers still sit in the raw data. That same block of scores also decides something else entirely: whether a country counts as an electoral democracy at all.

• 4. Electoral Versus Liberal Democracy

  ›

  A country needs a score of 7 or more out of 12 in political rights subcategory A. The report labels that subcategory Electoral Progress. Clearing it is one requirement for electoral democracy status. A country also needs an overall aggregate score of 20 or more in political rights. Since 2018, it has needed 30 or more in civil liberties as well. Territories are not eligible for this classification at all. Freedom House distinguishes electoral democracy from what it calls liberal democracy. Liberal democracy additionally requires a substantial array of civil liberties and a robust observance of democratic ideals. Most countries rated Free qualify as both electoral and liberal democracies. Some Partly Free countries qualify only as electoral, not liberal, democracies. A separate, narrower process decides which places even get counted as territories in the first place.

• 5. Countries And Territories

  ›

  The 2026 edition of the survey covers 195 'countries'. That figure includes the 193 United Nations member states plus Kosovo and Taiwan. It also separately rates 13 territories. Freedom House says it typically takes no position on territorial or separatist disputes as such. It focuses instead on political rights and civil liberties within a given geographical area. Officially, five factors decide whether a place earns its own territorial assessment. A territory must have separate governance from its parent state. Its conditions must be unique enough that ratings would likely differ from 'the relevant country or countries.' There must be active pressure for a change in political status, whether toward 'autonomy, independence, or incorporation into another country.' The territory needs enough stability to assess a full year and to allow comparisons over future years. Finally, its size or political significance factors into the decision. What has proven far more contested is whether the measurement itself deserves the trust it receives.

• 6. The Bias Debate

  ›

  Kenneth A. Bollen concluded in 1986 that assorted criticisms suggested some nations 'may have been incorrectly rated on Gastil's measures.' He found that none of them had demonstrated 'a systematic bias in all the ratings.' Bollen described most of the evidence as anecdotal, covering only a relatively small number of cases. He called the question of a systematic or sporadic slant in the ratings 'an open question.'

  Mainwaring and colleagues checked the index against Latin American countries year by year in 2001. They found a Pearson's coefficient of 0.82 between it and their own democracy index. Correlations with the other indices they studied ranged between 0.80 and 0.86.

  Gastil answered critics directly in 1990. He argued that most complaints rested on 'opinions about Freedom House rather than detailed examination of survey ratings.' Diego Giannone later disputed that conclusion. Adam Przeworski pushed further, arguing that Freedom House's definition of freedom emphasized liberties over their actual exercise. He pointed out that Americans are free to form parties and vote in presidential elections, yet only about half of them do. He wrote that in the United States, 'the same two parties speak in a commercially sponsored unison.'

  Freedom House answered more recent charges of ideological bias with a 2010 statement. It said the organization 'does not maintain a culture-bound view of freedom.' It said its standards apply to every country and territory 'irrespective of geographical location, ethnic or religious composition, or level of economic development.'

• 7. Conceptual Stretching

  ›

  Political scientists Andrew T. Little and Anne Meng studied Freedom House and the Varieties of Democracy project, known as V-Dem. They argued both rely on subjective rather than objective measures. This reliance, they concluded, leaves both projects open to coder bias.

  Diego Giannone borrowed a term from Giovanni Sartori to describe a different problem: 'conceptual stretching.' Citing a 2003 study by Landman and Hausermann, Giannone made a specific charge. Freedom House's index, he said, has been used as a tool for measuring democracy, good governance, and human rights all at once. He warned that 'an instrument used to measure everything, in the end, is not able to discriminate against anything.'

  Giannone also cited a 1981 conclusion by Scoble and Wiseberg about how the index adds up its parts. A civil liberty score of 4 combined with a political liberty score of 2 produces one total. A civil liberty score of 2 combined with a political liberty score of 4 produces the exact same total. Scoble and Wiseberg argued that 'the substantive interpretation of these different combinations is different.'

  Munck and Verkuilen went further still in 2002. They concluded that 'the aggregate data offered by Freedom House has to be accepted largely on faith.' No coding rules are published. Sources are not identified with precision. Disaggregated data has never been made available to independent scholars. Those complaints target how a single year gets measured. A separate strand of criticism asks whether the measurements can even be compared across years at all.

• 8. An Uncertain Time Series

  ›

  Diego Giannone published a paper in 2010 titled 'Political and ideological aspects in the measurement of democracy: the Freedom House case.' The paper reviewed changes to the survey's methodology going back to 1990. Giannone concluded that those shifts, tangled up with political considerations, mean the data 'do not offer an unbroken and politically neutral time series.' He argued they 'should not be used for cross-time analyses even for the development of first hypotheses.' The internal consistency of the data series, he wrote, remains 'open to question.'

  Freedom House's own website answers that critique directly. It says the organization has 'made a number of modest methodological changes to adapt to evolving ideas about political rights and civil liberties.' It also says the time series data 'are not revised retroactively.' Any changes, it adds, 'are introduced incrementally in order to ensure the comparability of the ratings from year to year.'

  Most of the retired territories were first added in the 1978 report, covering 1977. They received their last coverage in the 2000 report for that same year. Eastern Donbas is a more recent case. Freedom House rated it Not Free from 2020 until 2024, then stopped tracking it as a separate entry, one more line in a list of places the survey has picked up and later let go.