Childhood socioeconomic position and adolescent psychological outcomes: cross-cohort comparisons using data from three UK birth cohort studies

Abstract

Socioeconomic position [SEP] is a multi-dimensional and evolving construct, yet research rarely considers the issues presented by this complexity. Nonetheless, early socioeconomic disadvantage is associated with numerous consequences for young people,such as mental health difficulties, antisocial behaviour, and substance use, amongst others. Recent research indicates that the construct of SEP and these emotional and behavioural outcomes, and the relationship between them, are changing over time; however, these analyses are limited due to a lack of comparable data between cohorts.

Using an innovative methodological approach, retrospective harmonization, this thesis made data from three UK birth cohorts from the 1970s [BCS70], 1990s [ALSPAC], and 2000s [MCS] directly comparable. This thesis aimed to model how SEP has changed over the past 50 years and to explore how the relationships between SEP and adolescent emotional and behavioural outcomes have changed across cohorts using harmonized data. First, the extensive literature on the relationship between childhood SEP and mental health, antisocial behaviour, and substance use across the lifespan is reviewed (Chapter 2). This thesis thenexplores how to model SEP in a robust, theoretically-informed way and illustrate how it has changed over time using latent class analysis (Chapter 4). Following this, the sociodemographic characteristics of SEP latent classes were examined to determine who ismost at risk of disadvantage, and how their risk developed over time (Chapter 5). Next, therelationships between SEP latent classes and adolescent outcomes across cohorts were explored, including mental health (Chapter 6), antisocial behaviour (Chapter 7), and substance use (Chapter 8). Finally, the SEP latent classes were evaluated against individual SEP indicators as predictors of adolescent mental health outcomes to determine the utility of using latent class analysis to analyse socioeconomic inequalities (Chapter 9).

Results suggest that patterns of SEP have changed significantly over time, and the latent class characterised by the greatest relative disadvantage in each cohort differed from each other. However, the disadvantaged class in all three cohorts were more likely to experience additional adversities such as single parenting, parental or child disability, parental smoking, and maternal mental distress. Certain sociodemographic characteristics, such as single parenting, were more strongly related to the disadvantaged class in the MCS compared to the older cohorts. Adolescents in the disadvantaged class also experienced greater mental health difficulties, particularly externalising symptoms. Importantly, an increase in the relationship between SEP and mental health difficulties was observed in the MCS cohort. The disadvantaged class also engaged in more antisocial behaviour and were more likely to smoke across cohorts. The relationship between SEP and antisocial behaviour appeared to change qualitatively between cohorts; however, socioeconomic inequalities in alcohol use were strongest in the MCS, whereas inequalities in smoking and illicit drug use were strongest in the ALSPAC. Results from Chapter 9 also indicate that SEP latent classes are stronger and more accurate predictors of mental health outcomes when three or more classes are identified, but housing tenure and social class were the next best individual indicators across cohorts.

Overall, this thesis highlights the importance and utility of modelling the multi-dimensionality of SEP. Moreover, while the appearance of socioeconomic disadvantage has changed over time, the effects have become more deleterious for young people’s mental health. Yet poor outcomes are not ubiquitously associated with the disadvantaged class; young people from advantaged backgrounds are at risk of drinking more frequently and heavily in recent years. Harmonization underlies these robust cross-cohort comparisons and should be used to further explore the patterns and mechanisms for these changes over time.

Date of Award	Sept 2025
Original language	English
Supervisor	Orla McBride (Supervisor), Jamie Murphy (Supervisor) & Eoin McElroy (Supervisor)