Support for democracy

Between 1995 and 1998, in its Wave 3, the World Values Survey asked adults in 3939 countriesEstimandcount of countriesWeightsnone (count)Populationcountries with every political-system item above the floor in Waves 3, 5 and 7PeriodWaves 3, 5 and 7SourceWorld Values Survey and EVS countries to rate ways of governing their country one at a time, so a respondent could call both a democratic political system and a strong leader good. Counting every country once, 89%88.5% percent of respondentsEstimandequal-weight mean across countries of the weighted share calling a democratic political system a good way of governing the countryInterval95% CI 86.1% to 90.8%n49,767Weightsweight (S017) within country; equal weight per country across countriesPopulationadults aged 18 or older in the 39 countries observed in Waves 3, 5 and 7, equal weight per countryPeriodWave 3 (1995 to 1998)SourceWorld Values Survey and EVS of adults called a democratic system good and 38%38.2% percent of respondentsEstimandequal-weight mean across countries of the weighted share calling a strong leader who does not bother with parliament and elections a good way of governing the countryInterval95% CI 32.9% to 43.4%n49,235Weightsweight (S017) within country; equal weight per country across countriesPopulationadults aged 18 or older in the 39 countries observed in Waves 3, 5 and 7, equal weight per countryPeriodWave 3 (1995 to 1998)SourceWorld Values Survey and EVS called good a strong leader who does not have to bother with parliament and elections. Because each option was rated on its own, the two shares can overlap. The other options were army rule and experts rather than government making the decisions, each rated very good, fairly good, fairly bad or very bad. The chapter follows the same countries through Wave 5, in 2004 to 2009, and Wave 7, in 2017 to 2023, so every comparison across waves holds the set of countries fixed.

A democratic system held steady and a strong leader gained

By 2017 to 2023 the share calling a democratic system good had changed by +1.2 (95% CI −1.0 to 3.4)+1.2 percentage pointsEstimandequal-weight mean across countries of the change in the weighted share calling a democratic political system a good way of governing the countryInterval95% CI −1.0 to 3.4n114,207Weightsweight (S017) within country; equal weight per country across countriesPopulationadults aged 18 or older in the 39 countries observed in Waves 3, 5 and 7, equal weight per countryPeriodWave 3 to Wave 7SourceWorld Values Survey and EVS pointsPercentage pointThe unit for the difference between two percentages. A change from one share to another is stated in points, not in percent, so that absolute and relative change are not confused., an interval that includes zero. The share calling a strong leader good rose by +9.9 (95% CI 4.9 to 14.9)+9.9 percentage pointsEstimandequal-weight mean across countries of the change in the weighted share calling a strong leader who does not bother with parliament and elections a good way of governing the countryInterval95% CI 4.9 to 14.9n112,906Weightsweight (S017) within country; equal weight per country across countriesPopulationadults aged 18 or older in the 39 countries observed in Waves 3, 5 and 7, equal weight per countryPeriodWave 3 to Wave 7SourceWorld Values Survey and EVS points, and the share calling army rule good rose on the same timing. For both, the change from Wave 3 to Wave 5 has an interval that includes zero and the change from Wave 5 to Wave 7 does not. The democratic-system share starts near the top of the scale, so it has less room to rise than to fall.

The chapter also sorts respondents who answered the democratic-system, strong-leader and army-rule items into three groups. They are democracy only (a democratic system good and neither alternative), democracy and an alternative (a democratic system good and at least one alternative), and those who do not call a democratic system good. The first group shrank by −10.4 (95% CI −15.8 to −5.0)−10.4 percentage pointsEstimandequal-weight mean across countries of the change in the weighted share calling a democratic system good and neither a strong leader nor army rule goodInterval95% CI −15.8 to −5.0n105,561Weightsweight (S017) within country; equal weight per country across countriesPopulationadults aged 18 or older in the 39 countries observed in Waves 3, 5 and 7, equal weight per country who answered the democratic-system, strong-leader and army itemsPeriodWave 3 to Wave 7SourceWorld Values Survey and EVS points, the second grew by a similar amount, and the third changed by less than sampling error (Table 3). Inside the set the rise in approval of a strong leader is common but not general, and the democratic-system share rose in some countries and fell in others (Table 2).

Two survey programs and a fixed set of countries

The data are the World Values Survey (WVS) and the European Values Study (EVS), pooled into seven waves. "Good" means fairly or very good. A country is in the fixed set when enough respondents answered every item in each of the three waves, so each wave compares the same countries. Albania meets the rule and is left out because its 1998 survey shows an implausible response pattern, and the set has no country in Africa or the Middle East. Each country's share is weightedSurvey weightsNumbers that scale each respondent so the sample matches the population on known traits such as age, race, education and region. A weighted average counts a respondent with a large weight for more than one with a small weight., and a figure for the set gives every country the same weight. Its intervalConfidence intervalA range around an estimate that would contain the true value in 95% of repeated samples if the sampling model held. The book reports 95% intervals unless it says otherwise. is clustered by country, so it carries the differences between countries and is wider than one counting only sampling error inside them. In most countries the Wave 3 survey gives every respondent the same weight, so Wave 3 shares are in effect unweighted; without weights the three changes keep their signs. The report gives every estimate with its interval and sample size.

The pooled change averages two groups with opposite signs

Every country of the set had a WVS survey in Wave 3. By Wave 7, 2525 countriesEstimandcount of countriesWeightsnone (count)Populationcountries of the fixed set whose survey program changed from WVS in Wave 3 to EVS in Wave 7PeriodWave 3 and Wave 7SourceWorld Values Survey and EVS of them, all in Europe and Central Asia, had an EVS survey instead, and none of the rest is in that region. Where the program stayed the same, support for a democratic system changed by −2.8 (95% CI −5.6 to 0.1)−2.8 percentage pointsEstimandequal-weight mean across countries of the change in the weighted share calling a democratic political system a good way of governing the countryInterval95% CI −5.6 to 0.1n39,829Weightsweight (S017) within country; equal weight per country across countriesPopulationadults aged 18 or older in the 14 countries whose survey program stayed the same between Wave 3 and Wave 7, equal weight per countryPeriodWave 3 to Wave 7SourceWorld Values Survey and EVS points and approval of a strong leader by +16.0 (95% CI 7.9 to 24.0)+16.0 percentage pointsEstimandequal-weight mean across countries of the change in the weighted share calling a strong leader who does not bother with parliament and elections a good way of governing the countryInterval95% CI 7.9 to 24.0n38,981Weightsweight (S017) within country; equal weight per country across countriesPopulationadults aged 18 or older in the 14 countries whose survey program stayed the same between Wave 3 and Wave 7, equal weight per countryPeriodWave 3 to Wave 7SourceWorld Values Survey and EVS. Where it switched, the changes are +3.5 (95% CI 0.7 to 6.2)+3.5 percentage pointsEstimandequal-weight mean across countries of the change in the weighted share calling a democratic political system a good way of governing the countryInterval95% CI 0.7 to 6.2n74,378Weightsweight (S017) within country; equal weight per country across countriesPopulationadults aged 18 or older in the 25 countries whose survey program switched between Wave 3 and Wave 7, equal weight per countryPeriodWave 3 to Wave 7SourceWorld Values Survey and EVS and +6.5 (95% CI 0.2 to 12.8)+6.5 percentage pointsEstimandequal-weight mean across countries of the change in the weighted share calling a strong leader who does not bother with parliament and elections a good way of governing the countryInterval95% CI 0.2 to 12.8n73,925Weightsweight (S017) within country; equal weight per country across countriesPopulationadults aged 18 or older in the 25 countries whose survey program switched between Wave 3 and Wave 7, equal weight per countryPeriodWave 3 to Wave 7SourceWorld Values Survey and EVS points. The pooled change for a democratic system therefore averages a group whose support drifted down with a group whose support rose, while approval of a strong leader rose in both. Because the switch and the region go together, the data cannot say whether the difference belongs to the survey or to the place (program views of Figure 1).

The United States moved differently on support

The share of Americans calling a democratic system good fell by −5.6 (95% CI −8.1 to −3.1)−5.6 percentage pointsEstimandweighted share in 2017 minus 1995; independent samples, standard errors combinedInterval95% CI −8.1 to −3.1n3,942Weightsweight (S017) within countryPopulationadults aged 18 or older in the United StatesPeriod1995 to 2017SourceWorld Values Survey and EVS points between 1995 and 2017, while the change in the other countries of the set has an interval that includes zero. Only South Korea, Colombia, India and Montenegro fell further. On army rule, Americans were less approving than the average country of the set in 1995 and level with it in 2017. On a strong leader the American change relative to the other countries is +3.4 (95% CI −2.7 to 9.4)+3.4 percentage pointsEstimandUnited States change minus the equal-weight mean change of the other countries (standard errors combined) in the share calling a strong leader who does not bother with parliament and elections a good way of governing the countryInterval95% CI −2.7 to 9.4n116,888Weightsweight (S017) within country; equal weight per country across countriesPopulationUnited States against the 38 other countries of the fixed setPeriodWave 3 to Wave 7SourceWorld Values Survey and EVS points, an interval that includes zero but permits a difference as large as the army-rule one. The data neither show nor rule out a larger American move on that item. The American surveys also changed modeSurvey modeHow respondents were interviewed: in person, by phone, on paper or online. Mode can change answers, so a change in mode can look like a change in opinion., from face-to-face interviews in 1995 to the web in 2006 and web and telephone in 2017, and part of the change could come from the mode.

A second comparison places the United States among all 9292 countriesEstimandcount of countriesWeightsnone (count)Populationcountries surveyed in Wave 7 with an effective sample at or above the floorPeriodWave 7 (2017 to 2023)SourceWorld Values Survey and EVS countries surveyed in Wave 7 with enough respondents, a set that differs from the fixed one. There the United States is below the median country on support for a democratic system and on approval of a strong leader. It is at the median on army rule and on the top rating for the importance of living in a democracy. The median country's strong-leader share is 1413.6 percentage pointsEstimandmedian across countries of the weighted share calling a strong leader who does not bother with parliament and elections a good way of governing the country minus the United States shareWeightsweight (S017) within countryPopulationmedian across the 92 countries surveyed in Wave 7 minus the United StatesPeriodWave 7 (2017 to 2023)SourceWorld Values Survey and EVS points above the American one. Within the fixed set, Americans were also below the average country on a strong leader in both 1995 and 2017.

The age gap on a strong leader closed

Foa and Mounk read lower importance ratings among people born since 1980 as a generational difference [1]; Voeten found flat trends in established democracies [2]. Repeated cross-sections cannot settle the question, because birth year equals survey year minus age: comparing one age group across waves mixes birth cohortCohortPeople born in the same period, such as the same decade. Comparing cohorts shows whether a difference comes from when people were born, as distinct from their age or the year they were surveyed. with period, and following one cohort mixes age with period [3]. The age comparisons use the countries of the set with enough respondents in every age and cohort cell. Young adults, aged 1818 years of ageEstimandlowest age of the young groupWeightsnone (definition)Populationdefinition used in the chapterPeriodnot applicableSourceWorld Values Survey and EVS to 3434 years of ageEstimandhighest age of the young groupWeightsnone (definition)Populationdefinition used in the chapterPeriodnot applicableSourceWorld Values Survey and EVS, approved of a strong leader less than adults aged 5555 years of ageEstimandlowest age of the older groupWeightsnone (definition)Populationdefinition used in the chapterPeriodnot applicableSourceWorld Values Survey and EVS or more did in Wave 3, and of army rule less too. By Wave 7 the young-minus-older difference for a strong leader was +0.4 (95% CI −1.9 to 2.6)+0.4 percentage pointsEstimandequal-weight mean across countries of the share calling a strong leader who does not bother with parliament and elections a good way of governing the country among adults aged 18 to 34 minus that among adults aged 55 or olderInterval95% CI −1.9 to 2.6n51,421Weightsweight (S017) within country; equal weight per country across countriesPopulationadults aged 18 or older in the 30 countries with every age-group and cohort cell above the floor, equal weight per countryPeriodWave 7 (2017 to 2023)SourceWorld Values Survey and EVS points, an interval that includes zero, while the difference for a democratic system stayed close to zero in both waves. In the United States, with the mode caveat, young Americans in 2017 were more approving of army rule than older Americans, a gap that had been near zero in 1995 (United States view of Figure 3).

A gap between age groups that closes resembles a difference between cohorts, and a rise that reaches every cohort resembles a period change. Followed across waves, every birth cohort rises (Table 6). The data do not decide between the two readings, and both can hold. The importance question differs: in both waves that ask it, young adults give the top rating less often than older adults, and the change in that gap has an interval that includes zero.

What these surveys cannot tell you

The questions ask whether four ways of governing would be good, not whether a respondent would act on the view, and the strong-leader item may mean different things to different respondents. Chapter 5.4, Political violence and democratic norms has items that come closer to behavior. Each wave interviews different people, so the chapter describes how distributions changed and not how individuals did, and the data end in 2023, before the 2024 United States election. The file records no interview mode, some Wave 7 surveys outside the United States used web or self-completion, and the program switch is confounded with region. Differences in how the surveys were run can therefore look like change. Nothing here is causal. The chapter does not say why approval of a strong leader rose or why young Americans differ.

Go deeper

• The report: every estimate with its interval and sample size.
• Tables: every figure of the report as a table.

Reading paths that include this chapter:

• Trust and democracy
• Outside the United States

Data sources

• World Values Survey and EVSWorld Values Survey and European Values Study, Waves 3 to 7 (1995 to 2023): political-system battery, importance of democracy, perceived democraticness, left-right placement, ageYears: 1981-2023 (7 unified waves) · Weights: weight (S017) within country; equal weight per country across countries · License: Free download with registration; users must cite the release.
• European Social SurveyEuropean Social Survey, Round 6 (2012 to 2013) and Round 10 (2020 to 2022): democracy evaluations and satisfaction with democracyYears: 2002-2023 (rounds 1 to 11; round 11 was fielded in 2023 and 2024) · Weights: weight (anweight) within country; equal weight per country across countries · License: Open access after free registration in the ESS Data Portal; follow the ESS data citation requirements.

All data sources

References

1. Foa, Roberto Stefan; Mounk, Yascha (2016). The Danger of Deconsolidation: The Democratic Disconnect. Journal of Democracy 27(3): 5-17. doi:10.1353/jod.2016.0049
2. Voeten, Erik (2017). Are People Really Turning Away from Democracy?. Journal of Democracy Web Exchange, Online Exchange on Democratic Deconsolidation. www.journalofdemocracy.org/wp-content/uploads/2018/12/Journal-of-Democracy-Web-Exchange-Voeten_0.pdf
3. Norris, Pippa (2017). Is Western Democracy Backsliding? Diagnosing the Risks. Journal of Democracy Web Exchange, Online Exchange on Democratic Deconsolidation (also HKS Faculty Research Working Paper RWP17-012). www.journalofdemocracy.org/wp-content/uploads/2018/12/Journal-of-Democracy-Web-Exchange-Norris_0.pdf

How to cite

@misc{singh2026supportfordemocracy,
  author = {Singh, Vishal},
  title = {Support for democracy},
  year = {2026},
  url = {https://vishalsingh.org/politics/support-for-democracy},
  note = {In The Public Mind. Published 3 October 2026.}
}