To generate a broad-ranging picture of global views on China, this project brings together in a single dataset more than 2,500 survey results from 160 countries. It builds from key multi-country survey projects, including Pew Research Center, Asian Barometer, Afrobarometer, Latinobarómetro, Americas Barometer, Central Asia Barometer, Arab Barometer, Eurobarometer, Sinophone Borderlands, Democracy Perception Index, BBC World Service Globescan, Gallup World Poll, YouGov, IPSOS, and Transatlantic Trends, supplementing these with results from other multi-country and single-country surveys from, among other sources, AsiaBarometer, the Asian Research Network, Cabinet Office of Japan, Washington Institute, Genron NPO, World Public Opinion, Center for Insights in Survey Research, and Zogby Research Services.

The project team cast a wide net, searching for published data in languages including Chinese, Russian, Spanish, and Arabic, and identified well over 3,000 surveys from which the Global Public Opinion on China (GPOC) dataset was distilled.

The “Global Average” figure shown in the top right of the main page is calculated from all surveys taken in the given year, weighted by population, and reflects the percentage of the world’s population expressing positive, negative, and neutral views of China in surveys taken that year.

The various surveys are integrated into the global dataset via a process of careful selection and moderation. The first step is to account for differences in how questions are worded by selecting only those where positive-leaning and negative-leaning responses can be clearly differentiated. The most commonly selected questions asked respondents for opinions, feelings, or attitudes toward China, or their view of China’s influence in their country, region, or the world. 

The second step is to make the results comparable by calculating each survey’s percentage of positive, negative, and neutral responses, which flattens the effects of different measurement scales. This enables calculation of “net favorability” figures from the percentage of positive responses minus the percentage of negative responses — akin to China’s “approval rating.”

Third, a moderation formula is applied to identify and offset possible biases. The formula compares each source’s individual net favorability results against all other results from the same country and year, producing an estimation of the direction and magnitude of each source’s bias relative to comparable surveys. The source-moderated figure is formed by adding or subtracting half of the survey source’s estimated bias.

The focus of the GPOC is on trends identifiable across multiple surveys. Individual data points should be interpreted with care, making reference to the survey’s source, mode, number of respondents, and how the questions were worded. The great majority of results (98%) are derived from weighted national or urban samples of 600 or more respondents, which typically yield a margin of error in the range of ±2%–5%. Survey modes include face-to-face, telephone, and online, while a small number of surveys (3%) used a combination of modes.