A good coverage of the intended population is one of the cornerstones of survey research, and the success of Internet surveys is dependent on how well the intended population is covered. Countries differ in Internet penetration, threatening the international comparability of web surveys. The Eurobarometer collects data for the European Community across member and applicant states. Each wave consists of face-to-face interviews with adults aged 18 +. Among the questions asked in several years are access to Internet. This provides us with coverage figures over time for each country. The Eurobarometer also asks 'core’ questions in each wave, e.g. political attitude questions. As survey mode in the Eurobarometer is held equal (face-to-face) and the same battery of questions is asked over time and countries, this enables us to investigate how coverage bias would influence the results if the data would have been collected using Internet surveys instead of face-to-face interviews. In other words, this gives us an indication of the coverage bias over time and across countries. These results have consequences for how successful Internet panels should be formed.