The quality of data generated by any survey is a function of three main elements. These are:
(i) the representativeness of the sample;
(ii) the completeness of responses (i.e., the extent to which there is missing data, or item-non-response); and
(iii) the accuracy of the data (which in turn is a function of the questionnaire design, the accuracy of the responses provided, and the accuracy with which responses are recorded, coded and entered into the database).
In addition, for the HILDA Survey, added concerns arise as a result of the need to match data, both across individuals within the same households and within individuals. This is a consequence of both the presence of multiple instruments and the desire to construct household-level variables.
This paper provides an assessment of the data obtained for the first wave of the HILDA Survey (conducted in 2001) against each of these criteria.