JOURNAL OF CHILDREN AND MEDIA, cilt.7, ss.133-150, 2013 (SSCI)
Survey researchers confront a variety of problems when interviewing children and these are compounded when the research includes thousands of children across multiple countries. This article analyses some of the pre-testing and cognitive interview data in the EU Kids Online study for the difficulties children had in answering questions related to risk, patterns of answers to the final questionnaire provided by the children, and the evidence for providing socially desirable answers to questions related to risk and sensitive information. The format of the administration of the survey in the field was also examined. Age, but not gender, differences were found in understanding, response rates, levels of completion for an open-ended question, and in social desirability influences on answers. The format of administration (CAPI versus PAPI) made no difference in responses to sensitive or risk behaviour-related questions, but the presence of others in the room was related to less socially desirable responses. The research contributes to the methodological considerations of survey research with children.