Research

My work focuses on three broad areas that approach a shared question: Who decides what counts as a policy problem, and what counts as the right response?

01 Defining problems: People, place, inequality

How inequality is experienced, defined, and made visible across different contexts, including education in Ireland, migration and education in Ethiopia, and disadvantage in the UK. This work focuses on which communities become visible in policy, how problems come to be defined, and how policies reach those they are intended to serve. A common thread is the gap between how systems define problems and how they are experienced in practice.

Selected work

Longitudinal fieldwork on how girls navigate domestic labour, family expectations, and institutional gatekeeping.
How learning loss was defined and measured, and what this obscured about students' lived experiences, particularly in the Global South.
02 Institutions, evidence, and decision-making

How evidence is produced, governed, and used within large-scale systems, from donor-funded reform programmes in Ethiopia to national policy environments across the UK. This work examines why evidence often struggles to inform decisions, and what institutional conditions support more effective use. It highlights how governance arrangements, funding structures, and political contexts shape what counts as evidence and how it is used.

Selected work

How accountability systems tied to measurable outcomes reshape what counts as success.
How research relationships shape what evidence is produced and whose perspectives are prioritised.
03 Measurement, methods, and AI

How measurement shapes what is visible in policy and research, and how analytical choices determine what counts as evidence. This work develops and applies methods across contexts, from psychometric validation and culturally grounded measurement to NLP approaches, with a focus on how data and AI systems are increasingly embedded in decision-making. It emphasises the need for methods that are both technically robust and attentive to the assumptions they encode.

Selected work

Developing methods that start from local constructs rather than adapting existing instruments.
NLP approaches to analysing policy and research systems
Designing and building NLP pipelines for analysing large-scale policy data, making model assumptions visible in decision-making contexts. See AI projects and tools
View full publications