Response rates in business surveys are low, which implies a huge risk of selection bias. Usually no attempt is made to assess the extent of selection bias and published survey results might, therefore, not be a correct reflection of actual population characteristics. In this paper, it is argued that response rates cannot be improved to a sufficient degree and that assessment of selection bias is difficult in practice. It is concluded
that academic questionnaire surveys of businesses should be abandoned and that an alternative way of testing theory in populations must be found. Some of these alternatives are presented and discussed.