Choosing a national survey to investigate smoking behaviour: making comparisons between the General Household Survey, the British Household Panel Survey and the Health Survey for England.

Abstract

Smoking behaviour is of central concern to enhanced population health and there is a recognized need for detailed information concerning many different aspects of this behaviour. This paper assesses the utility of three large-scale national surveys according to their ability to provide such information. These are the General Household Survey (GHS), the British Household Panel Survey (BHPS) and the Health Survey for England (HSE). A number of important data items are identified and a comparative content analysis of the surveys is undertaken to indicate whether these items are present or absent in each source. Although current smoking status and consumption are covered in all three data sets, the GHS and the HSE provide information on many other aspects of the behaviour. Information on the major contextual variables can be found in each of the surveys. All sources can be used within a repeated cross-sectional analytical framework, and limited population-based longitudinal analyses can be undertaken with the GHS and the HSE. The BHPS theoretically opens up potential for a true longitudinal analysis of the cohort dynamics of individual smoking behaviour. Issues of place are best explored using the HSE, where the geographical information is most detailed. The paper highlights the importance of all three data sets as statistical modelling resources for investigating patterns of smoking prevalence. Although the advantages of the panel approach are noted for cohort-based longitudinal studies, there still remain some problems in undertaking these analyses because of the small number of BHPS waves carried out to date.