‘Long‐Termness’ with Regard to Sickness and Disability: An Example of the Value of Longitudinal Data for Testing Reliability and Validity

Abstract

Longitudinal data are used to explore the reliability of long‐term illness measures using information on ‘long‐term/permanent sickness or disability’ as recorded via the question on economic activity in the Census of England and Wales. Specifically it is asked ‘what evidence is there that individuals answer the item on long‐term sickness or disability in the same way at different time points?’. The research is based on data from the Office for National Statistics Longitudinal Study (LS) and the responses of persons who reported long‐term sickness or disability in 1971 are tracked to see if they also reported long‐term sickness or disability in 1981 and 1991. That a large number of people reported long‐term sickness or disability at three Census points suggests that the measure was reliable. In addition, it is suggested that a degree of content validity has been achieved with respect to the understanding of the concept of ‘long‐termness’. It is concluded that longitudinal data provide a valuable way of investigating the reliability and validity of frequently used categorical measures and can help to distinguish between the extent to which a phenomenon such as long‐term sickness is ‘new’ or ‘ongoing’.

Acknowledgements

This article draws on research funded by the Economic and Social Research Council (RES‐000‐22‐0081) under the Small Grants Scheme. The authors gratefully acknowledge the Office for National Statistics for access to the Longitudinal Study and in particular the contribution of Alec Ross. They would also like to thank the two anonymous referees for their helpful and constructive comments.

Notes

[1] A note on terminology. Two Census measures of illness/disability are discussed in this paper: sickness/disability and limiting long‐term illness (LLTI). When referring to the Census measures we use the terms ‘sickness’ and ‘LLTI’ respectively, but in general discussion of the substantive and methodological literature we use the more conventional term ‘illness’ to mean both.

[2] Standard Census testing was used, whereby respondents completed the test forms and their answers were verified through interview.