Item nonresponse is widely considered an important indicator of data quality. It decreases the available sample size for analyses and bears the risk of biased results if the missingness is not at random. The present study investigates item nonresponse rates in the first three rounds of the European Social Survey (ESS). The ESS is a biennial cross-national survey of attitudes and behaviours, first fielded in 2002. The authors' focus is on the average level of item nonresponse across 75 questions being part of the 'core modul' of the ESS questionnaire (cumulative item nonresponse). They describe the average amount of item nonresponse for all countries separately for the different types of item nonresponse: 'don't know', 'refusal', and 'no answer'. In addition they analyse the potential reasons for differences across countries in the main type of item nonresponse, i.e. 'don't know' nonresponse. This is done by multi-level modelling; the three hierarchical levels of respondents, interviewers, and countries are distinguished.