In this article the problem of survey non-response is examined with special reference to probability sampling in the Czech Republic. Non-response rates among Czech respondents in ISSP surveys between 1995 and 2005 were almost twice the rate recorded between 1991 and 1995 (25%). Such trends point to a decline in the 'survey climate'. While non-contacts and refusals in surveys are a significant problem, issues relating to how fieldwork is undertaken are equally important. The large fluctuations in non-contact rates and the relative success of the Czech Statistical Office in attenuating non-response rates demonstrates that prudent surveying strategies can be effective. An examination of two waves of the European Social Survey (ESS) reveals both the problems and potential strategies available for response rate enhancement. In this respect, all survey designers face the dilemma of balancing the benefits of data accuracy with increasing logistical costs. Improvement in survey quality necessitates consideration of many issues and the ability to make sensible trade-offs between competing research objectives.