Some Methodological Issues for the 2002 NATSISS

In this paper we introduce the particular issues involved in analysing the 2002 NATSISS. We discuss a number of aspects of the survey methodology including the scope, sample design and interviewing techniques. We pay particular attention to the different survey methodology used in Community Areas and Non Community Areas and the implications for analysis of the data, particularly that which uses the Confidentialised Unit Record File. A number of issues to do with sampling and non-sampling error are outlined, including how users can take into account sampling and non-sampling error when making conclusions. In particular, we give three examples of potential non-sampling error that show how users need to be aware of questionnaire design, interviewing techniques and imputation strategy when using the NATSISS.