Americans were more likely to name China, Russia and a wider range of countries as global superpowers when asked directly than when first asked whether any superpowers exist, a Pew Research Center survey experiment found.
The experiment, published Oct. 27, tested two ways to ask U.S. adults about countries that are both economically and militarily dominant. The survey ran March 23-29, 2026, on Pew’s nationally representative American Trends Panel.
In the direct version, 44% named China, compared with 32% in the two-step version, a 12-percentage-point difference. Russia was named by 24% and 14%, respectively. The United States was the most commonly named power in both formats, at 57% and 50%.
The findings concern how people answer questions about global power. They are measures of respondents’ views, rather than an assessment of the countries’ actual economic or military strength.
A preliminary question brought more uncertainty
The direct, or unfiltered, approach let people name up to six countries and offered options to say none or express uncertainty. The filtered approach first asked whether any global superpowers exist; people answering yes were then asked to name them.
Both versions produced similar overall patterns, with the United States followed by China and Russia as the most frequently named. But the filtered format produced more uncertainty: 23% expressed uncertainty at the first question and another 5% at the follow-up, for a combined 28%, compared with 18% in the direct format.
Respondents named 47 different countries in the unfiltered version and 30 in the filtered version. Pew researchers said the narrower range may be connected to the greater share of uncertain responses in the two-step approach.
The United Kingdom was named by 8% in the direct version and 3% in the filtered version; the European Union was named by 5% and 2%. Shares saying there were no superpowers were 11% and 14%. Pew described the differences between formats in its country-response chart as statistically significant.
Across all respondents in the experiment, 27% named three or more superpowers in the direct version, compared with 15% in the filtered version. The share naming one was 27% in both formats, while 16% named two in the direct format and 15% did so in the filtered format.
Interviews helped researchers replace confusing terms
The experiment grew from cognitive interviews with 12 U.S. adults in winter 2024. Interviewers asked follow-up questions to understand how participants interpreted the wording and arrived at answers.
Terms for a world with one, two or several superpowers confused some people. Participants more readily named countries than counted superpowers, and a couple initially interpreted the word as referring to supernatural abilities.
Those interviews led researchers to define a superpower in terms of both military and economic dominance and let people name multiple countries. The new formats also allowed respondents to say that no country qualified or that they were unsure, options missing from Pew’s older single-choice questions.
Previous answers reveal another format effect
Pew also examined 1,770 respondents who answered an experimental question in 2026 and its longstanding economic and military power questions in 2025. Their answers let researchers compare how people who had previously chosen one leader responded when they could name several.
Among people who previously placed the United States first on both measures, 44% named two or more superpowers in the direct format, compared with 29% in the filtered format.
Among those who previously split the leading economic and military positions between the United States and China, 57% named multiple superpowers with the direct question, compared with 30% with the filtered question. In that group, the filtered version more often produced answers of one superpower, 32% versus 18%, or none, 17% versus 9%.
The older questions asked people to choose a single leader from specified lists, so they are not directly comparable to the new open-ended questions. Views may also have changed between 2025 and 2026, Pew noted.
The American Trends Panel recruits U.S. adults through probability sampling and uses weighting to account for selection and nonresponse. Pew said it is most likely to use the simpler, direct version in future cross-national surveys because substantive results were broadly similar and the direct version was easier to administer and produced less nonresponse.
FingerLakes1.com is the region’s leading all-digital news publication. The company was founded in 1998 and has been keeping residents informed for more than two decades. Have a lead? Send it to news@fingerlakes1.com.
