pathway: Advanced Quantitative Methods

There is debate around the potential of social media to negatively impact wellbeing, particularly in young people. However, the literature base on this topic is inconclusive. My research aims to explore this potential association, particularly by investigating individual differences which may modulate the impact of social media. By triangulating evidence from self-report and passive measures of social media use, neurocognitive assessments, and both observational and experimental design, I hope to build a more nuanced view of whether specific patterns of use and/or cognitive biases predispose people to experiencing detrimental effects of social media.

Exposure to childhood trauma is a key social determinant of health and mental health. However, while research with adults has evidenced strong associations between childhood trauma and psychopathology, research with children and young people (CYP) is limited and has yielded mixed findings. Notably, the effects of trauma on CYP in low- and middle-income countries (LMIC) has been critically understudied despite disproportionally higher rates of trauma exposure compared to CYP in high-income countries. The overall aim of my PhD is to therefore provide systematic investigation of the effects of trauma exposure on the psychological and physiological health of CYP from LMICs.

Government policies are critical in contouring patterns of health and mitigating health inequalities. Previous literature showed how such policies may determine health outcomes. However, there is a scarcity of studies investigating health benefits of government interventions in different sectors than those directly related to health. My research aims to fill this gap by longitudinally examining the effects of public health and non-health policies on individual health indicators. I utilise multilevel SEM to uncover time-prolonged effects with bidirectional relationships between a person’s socioeconomic status and health and employ the Understanding Society dataset matched with administrative records.

My research focusses on methods of modelling voting patterns and social and political attitudes at granular levels, particularly in smaller geographical units than electoral districts. I hope to both expand the use of common techniques such as Multilevel Regression and Poststratification (MRP) and consider diverse alternative approaches such as the use of visual data with computer vision. These techniques can improve understanding of British electoral geography and the underlying social phenomenon which drive it, as well as considering electoral geography from the subjects’ perspectives – for example through the visual elements of street-scenes.

My project aims to extend the existing computational methods in understanding extremist language. I plan to use the state-of-the art transfer learning techniques in developing a Named Entity Recognition (NER) framework to predict the in-group and out-group concept of social psychology in the extremist communication. I plan to develop a model to detect actors across multiple extremist groups.

My research will centre around analysis of data from the English context to provide new evidence on the interaction between relative school starting age (those youngest and oldest in the school year) and early years ‘schoolification’, the long-term consequences for children and the mechanisms that underlie any effects.

I hope to explore associations between relative school starting age and children’s emotional development in Reception class, as well as discover whether discrepancies exist between parents and teachers’ perceptions and clinical diagnosis. Then, I will investigate how relative school starting age may be associated with trajectories of academic and socioemotional development into mid-to-late-adolescence and how these trajectories may vary according to indicators of emotional development in the Reception year, gender, maternal education and birth cohort.

Population ageing places significant pressures on health and social care services. Physical activity is one of the most promising avenues for reducing the socioeconomic burden of age-related diseases. Although mobile technologies could transform the healthcare industry, digital exclusion is stratified by socioeconomic status, emulating physical activity divides. My research will use methods such as multilevel modelling and machine learning to inform the development of a “just-in-time” adaptive intervention for older adults of low socioeconomic status. By monitoring dynamic biocultural variables (e.g. mood, physical/social environment), this digital intervention will aim to deliver personalised physical activity support when users need it most.

My research looks at the application of Advanced Quantitative Methods to support scalable democratic participation and policy making that incorporates wellbeing and sustainability outcome tracking and goal setting. Methodologically I am applying Natural Language Processing (NLP), probabilistic and graph-based techniques in the development of a computational framework that spans argument mining, policy analysis and econometrics. My work is inspired by two normative drivers: strengthening democracy through increasing citizen agency, and supporting fairness and sustainability as per the ‘Beyond GDP’ agenda.

A growing interdisciplinary literature examines the role of industry actors and conservative think-tanks in shaping the climate change debate and promoting inaction by manufacturing doubt about anthropogenic climate change. The aim of my research is to quantify the influence of these sceptics on the political discourse with a text-mining approach. The language of the sceptics is to be identified by analysing not only documents of sceptics but also of the environmental movement. Distinguishing between the two discourses will allow a nuanced textual exploration of political agendas and provide a framework for measuring the impact of each group on the political debate.