In China, the survey was conducted May 21-31, 2005 among a probability sample of 2,191 respondents in six major cities and their surrounding rural areas (Beijing, Chengdu, Guangzhou, Shanghai, Shenyang, and Wuhan). The sample is thus disproportionately urban and is not representative of the entire country. Interviews were conducted in-person, in the appropriate Chinese dialect, with adults ages 18-60. The margin of sampling error for the results is plus or minus 2%.