What impact do media have on public opinion when they are tightly controlled by the government, as in the People's Republic of China? Little reliable information is available for an answer, especially at the individual level. This study is a secondary analysis of a stratified random sample of 870 Beijing residents concerning economic reform in China during the mid-1980s. They were questionnaire-interviewed in 1986 with a 90 percent response rate. Regression and path analysis suggest that the government media were effective in increasing knowledge about reform and in affecting some attitudes, but it was less effective or totally ineffective in others. When compared with U.S. media studies, the magnitude of effects appears larger, a result consistent with message competition theory. This relatively new theory has been offered as an alternative to the classic 'selective exposure' and 'two-step flow' theories in explaining why little correlation has been found between individual media use and attitudes in the West. The theory predicts that competing messages tend to neutralize net effects of the media, while controlled media systems tend to show increased net effects. © 1994 World Association for Public Opinion Research.