National statistical institutes in many countries are faced with budget constraints on the one hand and demands for more and more detailed information on the other. Often CAPI and CATI surveys are conducted for data collection. These modes of data collection are expensive. At first sight, web surveys seem to be an interesting and attractive means of data collection. They provide simple, cheap and fast access to a large group of potential respondents.

The question arises whether web surveys can play a role in finding a solution of this dilemma of official statistics. Web surveys may be fast and cheap, but there alse methodological challenges. Specific groups in the population are under-represented because they have lower access to Internet. Furthermore, recruitment of respondents is often based on self-selection. Both under-coverage and selfselection may lead to biased estimates.

The paper describes these methodological problems. It also explores the effect of various correction techniques (adjustment weighting and use of reference surveys). It is argued in that self-selection web surveys cannot play such a role. However, when web surveys are carried out within the framework of probability sampling, there are certainly possibilities, as a single-mode survey or as one of the modes in a mixed-mode survey.

Changing from a CAPI or CATI survey to a web survey also means to attention has to be paid to mode effects. There are no interviewers involved in a web survey, and this has consequences for the way in which respondents answer questions.

Web surveys are already applied as one of the modes in mixed-mode survey. Web surveys may be successful were other modes fail to achieve sufficient response. However, it should be realizde that mode effects may have a negative impact on the quality of the collected data. It is clear that more research is necessary to optimally deploy web surveys for official statistics.