Abstract
In the evolving landscape of data-driven development, understanding the impact of research on national and international policy is paramount. This paper delves into the challenging realm of Monitoring, Evaluation, and Learning (MEL) in research for development, particularly focusing on an underexplored area of data justice: tracking policy influence from scientific research.
This paper combines an extensive literature review with 10 in-depth expert interviews that systematize the importance of tracking research influence, main methodologies used, and the current characteristics of the data ecosystem in MEL for development. It then uses a computational data science approach to assess the influence of research on open government policies in Latin America. Specifically, it explores a searchable index of millions of policy documents (Overton), combined with contextual information, to assess possible metrics of research influence on policy.
While unveiling the potential of data and tools based on large language models (LLMs) in the MEL domain, the paper uncovers a stark reality: the data is incomplete, and its advantages are skewed towards the Global North. The coverage and representation in the Global South are significantly lacking, with idiosyncratic differences influencing how research is used in policy decisions.
This research contributes to the broader discourse on fairness and justice in data and computational systems. It emphasizes the need for nuanced approaches in evaluating research impact on policy decisions, advocating for a more equitable data landscape in the pursuit of global development goals.