- Course: PhD in Medicine
- PhD title: Statistical and simulation-based modelling approaches for causal inference in longitudinal data: Integrating counterfactual thinking into established methods for longitudinal data analysis
- Year of graduation: 2020
- Nationality: American
- Job title: Postdoctoral Research Fellow in Urban Analytics
- Company: University of Leeds
- LinkedIn: https://www.linkedin.com/in/kellynarnold/
Kellyn Arnold is a Postdoctoral Research Fellow who uses both statistical- simulation-based modelling as part of her interdisciplinary research. She is based in the Leeds Institute for Data Analytics (LIDA), where she collaborates with colleagues in the Schools of Medicine and Geography. She is co-supervised by Professor Allison Heppenstall, and researches how formal causal inference methods can be integrated into simulation-based modelling approaches.
Modelling ‘complex interactions’
Kellyn’s work investigates how these methods can be applied to refine the process of estimating ‘cause and effect’ scenarios in geographical and other complex structures. This can enable stakeholders to predict what the different outcomes might be in the future as the result of specific interventions. Her work considers some of the challenges associated with modelling complex behaviour patterns of people living and working in cities, which produce enormous amounts of data on a daily basis. Historically, these methods have only been applied within a narrow range of research areas.
“Formal statistical causal inference methods have been around for several decades, but their use has been largely confined to a few niche disciplines such as epidemiology and economics,” explained Kellyn.
“My work with Professor Heppenstall seeks to apply some of these methods within geographical contexts, which often involve hierarchical structures and complex interactions. In so doing, we hope that we can bring more robustness to the process of estimating causal effects within these contexts, and also understand some of the limitations of current methods for dealing with complexity.”
Studying for a PhD
Kellyn first began her studies at the University of Leeds with an MSc in Epidemiology and Biostatistics, from which she went on to study a PhD. Kellyn said:
“During my PhD I had the opportunity to do methodology research with my supervisor Professor Mark Gilthorpe. I was keen to continue that research at PhD level, so we developed a proposal together with my other supervisor Professor Alison Heppenstall and eventually won funding for it from the UK Economic and Social Research Council (ESRC).”
She continued: “I spend a lot of time researching methods and formal causal inference theory, and then look for ways these can be applied across different domains. On the practical side, this often involves data analysis and/or simulation studies, so I spend a lot of time at a computer writing code.”
My PhD was really a cross-disciplinary effort to bridge some of the concepts and methodology across two disparate research domains.
She added: “Both my supervisors are interested in causal inference but work in different research areas: medicine and geography, respectively, and with different methods. Therefore, my PhD was really a cross-disciplinary effort to bridge some of the concepts and methodology across two disparate research domains.”