How do children answer questions about frequencies and quantities? Evidence from a large-scale field test

There is still much we have to learn about the best ways of obtaining accurate and comprehensive information from children without undue burden. This paper describes the findings from two experiments undertaken to develop response categories to maximise data quality for a series of questions intended to be useable with 11 year olds. The experiments were part of the instrument development for the child self-completion questionnaire for the fifth (Age 11) Survey of the Millennium Cohort Study (MCS), a large-scale national birth cohort study. The aim was to ascertain how to get the best quality estimates of frequencies of activity or quantities of consumption from children in response to potentially sensitive questions in a self-completion context. Specifically we set out to compare more and less specific response categories of two different types: descriptive versus numerical frequency responses, and banded versus open ended quantitative questions. To this end, we conducted two randomised experiments to evaluate responses to questions on bullying and alcohol consumption with a large-scale sample of children aged 11-16. We find that how response options to questions are presented matters for children’s responses. Questions need to be clear to avoid ambiguity in interpretation. When asking questions about quantities, such as amount of alcoholic drinks consumed, we conclude that better quality data is provided by fixed category rather than open ended response options.