Our surveys internationally are conducted via telephone or face-to-face interviews, depending on the country. The results are based on national samples, unless otherwise noted. Included here is detailed information, such as mode of interview, sampling design, margin of error, and design effect, for each country we survey, organized by country and by year.

For more general information on how we conduct our international survey research, visit here.

The margin of sampling error reported is based on all interviews conducted in a country. The margin of error takes into account the design effect due to clustering and weighting, where applicable. For results based on the full sample in a given country, one can say with 95% confidence that the error attributable to sampling and other random effects is plus or minus the margin of error. The margin of error is larger for results based on subsamples in the survey. Sample sizes and sampling errors for subgroups are available upon request. In addition to sampling error, one should bear in mind that question wording and practical difficulties in conducting surveys can introduce error or bias into the findings of opinion polls.

SELECT A COUNTRY
AND/OR A YEAR



Survey Methods for all available countries in 2016

Country: Australia
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: Random Digit Dial (RDD) probability sample of landline households (50% of sample) stratified by region and cell phone users (50% of sample). Individuals within landline households are selected using the youngest adult male or female method. Interviews in the cell sample are conducted with the person who answered the phone, if age 18 or older and sole user of phone, or by using the youngest adult male or female method when two or more individuals use the same cell phone. For both landline and cell samples, up to 10 phone calls are made to complete the interview with the selected respondent.
Mode: Telephone
Languages: English
Fieldwork Dates: April 6 - 28, 2016
Sample Size: 1,000
Margin of Error: 4.1 percentage points
Representative: Adults 18 plus in telephone households
Primary Vendor: Princeton Survey Research Associates International
Weighting Variables: Gender, age, education, region, metro area, phone use and probability of selection of respondent
Design Effects: 1.71
Country: Canada
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: Random Digit Dial (RDD) probability sample of landline households (50% of sample) stratified by geographic area and of cell phone users (50% of sample). Individuals within landline households are selected using the youngest person method. Interviews in the cell sample are conducted with the person who answered the phone, if age 18 or older. For both landline and cell samples, at least 10 phone calls are made to complete the interview with the selected respondent.
Mode: Telephone
Languages: English, French
Fieldwork Dates: April 5 - 20, 2016
Sample Size: 1,020
Margin of Error: 3.9 percentage points
Representative: Adults 18 plus in telephone households (excluding residents of Yukon, Nunavut and Northwest Territories)
Primary Vendor: Princeton Survey Research Associates International
Weighting Variables: Gender, age, education, region, urbanity and probability of selection of respondent
Design Effects: 1.63
Country: China
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: Multi-stage, area probability design. Primary sampling units (PSUs) are administrative units (districts, county-level cities and counties) stratified by regional-economic zones and type of administrative unit. Four provincial-level cities are selected with certainty. Within these cities, PSUs are grids within districts.
The number of effective PSUs is 71.
Mode: Face-to-face
Languages: Chinese (Mandarin, Shanghai, Pucheng, Longjiang, Ningxiang, Lanzhou, Pingyao, Yucheng, Qinhuangdao, Beijing, Tiandeng, Hangzhou, Fengjie, Lianghe, Kunming, Danfeng, Jiangmen, Guangzhou, Lushan, Foshan, Jingning, Yangzhou, Senhxian, Fuyang, Changzhi, Hexian, Tuanfeng, Yulin, Shulan, Longchang, Siping, Gaoping, Nehe, Jiedong, Pingyi, Chongqing, Chengde, Tianchang, Jixian, Xiushan, Shaoxing, Yongcheng, Liuyang, Xingcheng, Shaoyang, Ningdou, Xingping, Cenxi, Tianjin, Dengta, Nankang, Zibo, Hancheng, Huaying, Bazhou, Pizhou, Jinghong, Xingning, Lufeng, Wenling, Ruian, Yuyao, Qufu, Liyang, Changle, Gaoan, Jiangyin dialects)
Fieldwork Dates: April 6 - May 8, 2016
Sample Size: 3,154
Margin of Error: 3.7 percentage points
Representative: Adult population 18 plus (excluding Tibet, Xinjiang, Hong Kong and Macao)
Primary Vendor: Princeton Survey Research Associates International (Note: The results cited are from Horizonkey’s self-sponsored survey.)
Weighting Variables: Gender, age, region, urbanity and probability of selection of respondent
Design Effects: 4.45
Country: France
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: List-assisted Random Digit Dial (RDD) probability sample of landline households (50% of sample) stratified by region (NUTS1) and urbanity (urban, semi-urban, semi-rural, rural), and Random Digit Dial (RDD) probability sample of cell phone users (50% of sample) stratified by mobile network operators and number stem prefix. Individuals within landline households are selected using the Rizzo method. Interviews in the cell sample are conducted with the person who answered the phone, if age 18 or older. For both landline and cell samples, at least 10 phone calls are made to complete the interview with the selected respondent.  
Mode: Telephone
Languages: French
Fieldwork Dates: April 5 - 23, 2016
Sample Size: 999
Margin of Error: 4.3 percentage points
Representative: Adults 18 plus in telephone households
Primary Vendor: TNS BMRB
Weighting Variables: Gender, age, education, region and probability of selection of respondent
Design Effects: 1.9
Country: Germany
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: List-assisted Random Digit Dial (RDD) probability sample of landline households (70% of sample) stratified by region (NUTS3), and Random Digit Dial (RDD) probability sample of cell phone users (30% of sample) stratified by mobile network operators. Individuals within landline households are selected using the Rizzo method. Interviews in the cell sample are conducted with the person who answered the phone, if age 18 or older. For both landline and cell samples, at least 10 phone calls are made to complete the interview with the selected respondent. 
Mode: Telephone
Languages: German
Fieldwork Dates: April 4 - 23, 2016
Sample Size: 1,000
Margin of Error: 4.1 percentage points
Representative: Adults 18 plus in telephone households
Primary Vendor: TNS BMRB
Weighting Variables: Gender, age, education, region and probability of selection of respondent
Design Effects: 1.71
Country: Greece
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: Multi-stage, area probability design. Primary sampling units (PSUs) are municipalities (2011 LAU1; nomoi or ‘settlements’) stratified by region (NUTS1) and urbanity (DEGURBA). Eight urban municipalities are selected with certainty. Within these urban municipalities, the PSUs are square kilometer grids. The number of effective PSUs is 100. Individuals within households are selected using the Kish method. At least four attempts are made to complete the interview with the selected respondent. 
Mode: Face-to-face
Languages: Greek
Fieldwork Dates: April 14 - May 12, 2016
Sample Size: 1,007
Margin of Error: 4.7 percentage points
Representative: Adult population 18 plus (excluding the least densely populated PSUs; and the islands in the Aegean and Ionian Seas)
Primary Vendor: TNS BMRB
Weighting Variables: Gender, age, education, region, urbanity and probability of selection of respondent
Design Effects: 2.35
Country: Hungary
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: Multi-stage, area probability design. Primary sampling units (PSUs) are towns and villages (LAU2; városok, megyei jogú városok, nagyközségek, községek) stratified by region (NUTS1) and urbanity (DEGURBA). Thirteen urban municipalities are selected with certainty. Within these urban municipalities, the PSUs are square kilometer grids. The number of effective PSUs is 100. Individuals within households are selected using the Kish method. At least four attempts are made to complete the interview with the selected respondent.
Mode: Face-to-face
Languages: Hungarian
Fieldwork Dates: April 16 - May 10, 2016
Sample Size: 1,005
Margin of Error: 4.5 percentage points
Representative: Adult population 18 plus (excluding the least densely populated PSUs)
Primary Vendor: TNS BMRB
Weighting Variables: Gender, age, education, region, urbanity and probability of selection of respondent
Design Effects: 2.14
Country: India
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: Multi-stage, area probability design. Primary sampling units (PSUs) are urban settlements and rural districts stratified by region and urbanity. Nine large cities are selected with certainty. Within these large cities, PSUs are sub-grids. The number of effective PSUs is 167. Individuals within households are selected using a Kish grid and gender quotas based on national figures. At least three attempts are made to complete the interview with the selected respondent. The sample is disproportionately urban, but the data are weighted to reflect the actual urban/rural distribution in India.
Mode: Face-to-face
Languages: Hindi, Bengali, Tamil, Telugu, Kannada, Gujarati, Marathi, Odia
Fieldwork Dates: April 7 - May 24, 2016
Sample Size: 2,464
Margin of Error: 3.2 percentage points
Representative: Adult population 18 plus (conducted in the Union Territory of Delhi and in 15 of the 17 most populous states, Kerala and Assam excluded)
Primary Vendor: Princeton Survey Research Associates International
Weighting Variables: Gender, age, education, region, urbanity and probability of selection of respondent
Design Effects: 2.61
Country: Italy
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: Multi-stage, area probability design. Primary sampling units (PSUs) are municipalities (LAU2; comuni or ‘local councils’) stratified by region (NUTS1) and urbanity (DEGURBA). Five urban municipalities are selected with certainty. Within these urban municipalities, the PSUs are square kilometer grids. The number of effective PSUs is 100. Individuals within households are selected using the Kish method. At least four attempts are made to complete the interview with the selected respondent. 
Mode: Face-to-face
Languages: Italian
Fieldwork Dates: April 9 - May 10, 2016
Sample Size: 1,016
Margin of Error: 4.6 percentage points
Representative: Adult population 18 plus (excluding the least densely populated PSUs)
Primary Vendor: TNS BMRB
Weighting Variables: Gender, age, education, region, urbanity and probability of selection of respondent
Design Effects: 2.25
Country: Japan
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: List-assisted Random Digit Dial (RDD) probability sample of landline households (70% of sample) stratified by prefecture and Random Digit Dial (RDD) probability sample of cell phone users (30% of sample) stratified by mobile network operators. Individuals within landline households are selected using the Rizzo method. Interviews in the cell sample are conducted with the person who answered the phone, if age 18 or older. For both landline and cell samples, at least 5 phone calls are made to complete the interview with the selected respondent. 
Mode: Telephone
Languages: Japanese
Fieldwork Dates: April 26 - May 29, 2016
Sample Size: 1,000
Margin of Error: 4.1 percentage points
Representative: Adults 18 plus in telephone households
Primary Vendor: TNS BMRB
Weighting Variables: Gender, age, education, region and probability of selection of respondent
Design Effects: 1.77
Country: Kenya
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: Multi-stage, area probability design. Primary sampling units (PSUs) are sub-locations stratified by province and urbanity. The number of effective PSUs is 125. Individuals within households are selected using a Kish grid. At least three attempts are made to complete the interview with the selected respondent.
Mode: Face-to-face
Languages: Swahili, English
Fieldwork Dates: March 29 - April 19, 2016
Sample Size: 1,124
Margin of Error: 4.7 percentage points
Representative: Adult population 18 plus (excluding some areas of Taita and Turkana due to heavy rains and security concerns and some areas of Mandera and Garissa due to an administrative error)
Primary Vendor: ORB International
Weighting Variables: Gender, age, education, region, urbanity and probability of selection of respondent
Design Effects: 2.53
Country: Netherlands
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: List-assisted Random Digit Dial (RDD) probability sample of landline households (70% of sample) stratified by municipality (LAU2), and Random Digit Dial (RDD) probability sample of cell phone users (30% of sample) stratified by mobile network operators. Individuals within landline households are selected using the Rizzo method. Interviews in the cell sample are conducted with the person who answered the phone, if age 18 or older. For both landline and cell samples, at least 10 phone calls are made to complete the interview with the selected respondent. 
Mode: Telephone
Languages: Dutch
Fieldwork Dates: April 4 - 25, 2016
Sample Size: 999
Margin of Error: 4.4 percentage points
Representative: Adults 18 plus in telephone households
Primary Vendor: TNS BMRB
Weighting Variables: Gender, age, education, region and probability of selection of respondent
Design Effects: 2.05
Country: Nigeria
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: Multi-stage, area probability design. Primary sampling units (PSUs) are primary localities stratified by region and urbanity. The number of effective PSUs is 125. Individuals within households are selected using a Kish grid and gender quotas based on national figures. At least three attempts are made to complete the interview with the selected respondent.
Mode: Face-to-face
Languages: English, Hausa, Yoruba, Igbo
Fieldwork Dates: March 31 - May 16, 2016
Sample Size: 1,094
Margin of Error: 3.8 percentage points
Representative: Adult population 18 plus (excluding some areas of Borno and Yobe due to security concerns)
Primary Vendor: ORB International
Weighting Variables: Gender, age, religion, region, urbanity and probability of selection of respondent
Design Effects: 1.61
Country: Poland
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: Multi-stage, area probability design. Primary sampling units (PSUs) are municipalities (LAU2; gminy) stratified by region (NUTS1) and urbanity (DEGURBA). Seven urban municipalities are selected with certainty. Within these urban municipalities, the PSUs are square kilometer grids. The number of effective PSUs is 100. Individuals within households are selected using the Kish method. At least four attempts are made to complete the interview with the selected respondent. 
Mode: Face-to-face
Languages: Polish
Fieldwork Dates: April 8 - May 4, 2016
Sample Size: 1,003
Margin of Error: 4.6 percentage points
Representative: Adult population 18 plus (excluding the least densely populated PSUs)
Primary Vendor: TNS BMRB
Weighting Variables: Gender, age, education, region and probability of selection of respondent
Design Effects: 2.22
Country: South Africa
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: Multi-stage, area probability design. Primary sampling units (PSUs) are sub-places stratified by province, urbanity and metropolitan area, and race. The number of effective PSUs is 131. Individuals within households are selected using a Kish grid and gender quotas based on national figures. At least three attempts are made to complete the interview with the selected respondent.
Mode: Face-to-face
Languages: English, Zulu, North Sotho, Afrikaans, South Sotho, Xhosa
Fieldwork Dates: June 8 - July 9, 2016
Sample Size: 1,112
Margin of Error: 4.6 percentage points
Representative: Adult population 18 plus
Primary Vendor: ORB International
Weighting Variables: Gender, age, education, region, urbanity, race and probability of selection of respondent
Design Effects: 2.41
Country: Spain
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: List-assisted Random Digit Dial (RDD) probability sample of landline households (50% of sample) stratified by region (NUTS2) and community size, and Random Digit Dial (RDD) probability sample of cell phone users (50% of sample) stratified by mobile network operators. Individuals within landline households are selected using the Rizzo method. Interviews in the cell sample are conducted with the person who answered the phone, if age 18 or older. For both landline and cell samples, at least 10 phone calls are made to complete the interview with the selected respondent. 
Mode: Telephone
Languages: Castilian Spanish
Fieldwork Dates: April 4 - 22, 2016
Sample Size: 1,002
Margin of Error: 4.1 percentage points
Representative: Adults 18 plus in telephone households
Primary Vendor: TNS BMRB
Weighting Variables: Gender, age, education, region and probability of selection of respondent
Design Effects: 1.76
Country: Sweden
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: List-assisted Random Digit Dial (RDD) probability sample of landline households (40% of sample) stratified by municipality (LAU2), and Random Digit Dial (RDD) probability sample of cell phone users (60% of sample) stratified by mobile network operators. Individuals within landline households are selected using the Rizzo method. Interviews in the cell sample are conducted with the person who answered the phone, if age 18 or older. For both landline and cell samples, at least 10 phone calls are made to complete the interview with the selected respondent. 
Mode: Telephone
Languages: Swedish
Fieldwork Dates: April 4 - 23, 2016
Sample Size: 1,000
Margin of Error: 4.3 percentage points
Representative: Adults 18 plus in telephone households
Primary Vendor: TNS BMRB
Weighting Variables: Gender, age, education, region and probability of selection of respondent
Design Effects: 1.94
Country: United Kingdom
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: List-assisted Random Digit Dial (RDD) probability sample of landline households (70% of sample) stratified by region (NUTS3), and Random Digit Dial (RDD) probability sample of cell phone users (30% of sample) stratified by mobile network operators. Individuals within landline households are selected using the Rizzo method. Interviews in the cell sample are conducted with the person who answered the phone, if age 18 or older. For both landline and cell samples, at least 10 phone calls are made to complete the interview with the selected respondent. The sample includes an oversample in the Scotland region (NUTS1), but the data are weighted to reflect the actual regional distribution in the UK.
Mode: Telephone
Languages: English
Fieldwork Dates: April 4 - May 1, 2016
Sample Size: 1,460
Margin of Error: 3.3 percentage points
Representative: Adults 18 plus in telephone households
Primary Vendor: TNS BMRB
Weighting Variables: Gender, age, education, region and probability of selection of respondent
Design Effects: 1.67
Country: United States
Year: 2016
Survey: Global Attitudes Spring 2016
Sample Design: Random Digit Dial (RDD) probability sample of landline households (25% of sample) and cell phone users (75% of sample) stratified by county. Individuals within landline households are selected using the youngest male or female method. Interviews in the cell sample conducted with the person who answered the phone, if age 18 or older. For both landline and cell samples, up to seven phone calls are made to complete the interview with the selected respondent.
Mode: Telephone
Languages: English, Spanish
Fieldwork Dates: April 4 - 24, 2016
Sample Size: 1,003
Margin of Error: 3.4 percentage points
Representative: Adults 18 plus in telephone households
Primary Vendor: Princeton Survey Research Associates International
Weighting Variables: Gender, age, education, race, Hispanic origin and nativity, region, population density, phone use and probability of selection of respondent
Design Effects: 1.21