摘要：Respondents' interpretations of two survey questions were examined by questioning them about the process by which they arrived at their answers and their understanding of key components of the questions. This process revealed a range of interpretations for each question and its constituent components, and some respondents chose to interpret each question in a way that allowed them to express a strongly held opinion on the subject (AIDS or the unemployed). These results re-emphasise that researchers should not automatically assume that the questions they ask are the ones respondents answer. Failure to recognise this fact may have serious implications for decision makers using survey results and for the emphasis being placed in research on data analysis rather than data quality
