Understanding variations in children’s subjective well-being: a longitudinal analysis -PhD thesis-
There is a growing research and policy interest in the topic of subjective well-being (SWB), including in relation to children. Quite a lot is now known, from cross-sectional analysis, about factors associated with variations in the SWB of children in the UK. However there is a lack of longitudinal analysis which can help to clarify the mechanisms involved. This is an important gap if research on children’s SWB is to generate findings which can be practically useful. In addition the issues of the conceptualisation and measurement of children’s SWB have not been fully explored in UK research. This dissertation sets out to address these evidence gaps. The analysis uses data from three panel and cohort studies – the British Household Panel Survey, the Understanding Society survey and the Millennium Cohort Study – all of which include large, and broadly representative, samples of children in the UK and ask children some questions about their SWB The dissertation begins by addressing conceptual and measurement issues. It explores the extent to which the data support the most common conceptual framework of SWB – the tripartite model – and identifies the most appropriate measures of SWB to use for the analysis. It then aims to answer two broad substantive research questions. First, to what extent are early childhood factors associated with later subjective well-being (at 11 years old)? Second, how does subjective well-being vary between the ages of 11 and 15 and what factors are associated with this variation? The analysis draws on Bronfenbrenner’s ecological model of child development to theorise the ways in which current and historical factors may affect children’s SWB. It considers a range of both contextual (e.g. family socio-economic status) and process (e.g. parent-child relationships) variables. The analytical methods used for this purpose are logistic regression; linear regression, including fixed effects and random effects models for panel data; and latent growth curve modelling. Additionally, confirmatory factor analysis is utilised to test measures of SWB. The key findings are as follows. First, in terms of conceptual and measurement issues, the data provide support for the tripartite model of SWB commonly used in the literature on adults’ SWB. The psychometric analysis also provides new insights into the way in which variables in the data sets might best be used to represent SWB. Second, children’s life satisfaction, happiness and sadness at the age of 11 are not strongly predicted by their family and socio-economic circumstances at nine months old, nor by the quality of their relationship with their main parent at three years old. Third, there is a much stronger relationship between contemporaneous factors such as family relationships and bullying and trajectories of children’s SWB between the ages of 11 and 15, with some evidence that these factors predict future as well as current SWB. These conclusions hold even when making use of the longitudinal nature of the data to control for unobserved fixed explanatory variables. Overall, the analysis presented suggests that how children feel about their lives can best be understood in terms of recent rather than historical factors and provides further evidence of the much stronger associations between children’s SWB and quality of family and peer relationships, rather than socio-economic factors. It also suggests that the etiology of SWB is different from that of other childhood well-being measures, and it is argued that this is a strength that can help to stimulate new directions for social research on children and childhood. The analysis, and the conclusions that can be drawn from it, are limited by the quality of SWB measures available and the timing of data collection for the studies. It will be valuable for future research specifically on children’s SWB, and new longitudinal and panel studies in the UK, to use better SWB measures. There is also a need for longitudinal research over shorter time periods to explore further the directions of association between SWB and other factors.
ORCA - Online Research @ Cardiff - http://orca.cf.ac.uk/111878/