Scale of Harm: Estimating child exploitation trafficking in the Philippines

I worked with the International Justice Mission (IJM) on their Scale of Harm project, which aimed to estimate how many children in the Philippines are being trafficked to produce child sexual exploitation material. Using survey data and statistical models, we estimated how common this crime is across different parts of the country. The results helped IJM and the Philippine government understand the true size of the problem and plan where to focus their efforts to protect children. The law in the Philippines and the US was changed as a result of this work.

Multiple Systems Estimation

When people are victims of crimes like modern slavery, most are never identified by any official organisation. Multiple systems estimation is a way of using the small number of victims who do appear in different databases to estimate how many are being missed. I develop new statistical and machine learning methods using neural networks to make these estimates more accurate. This work helps governments and charities understand the true scale of hidden crimes so they can respond more effectively.

Neural Methods for Multiple Systems Estimation Models

Preprint, January 2026. arXiv

Neural Methods for Multiple Systems Estimation

Julia package for neural posterior estimation for multiple systems estimation with censored or missing data. Uses simulation-based inference to estimate the size of hidden populations. GitHub

A Partially Pooled Network Scale-up Method Model: Detailed Estimation of Child Sexual Exploitation Material Trafficking Prevalence in Philippine Municipalities

Journal of the Royal Statistical Society Series C, 2025. Paper

Cognitive Impairment and Exploitation: Connecting Fragments of a Bigger Picture through Data

Abubakar, A., Seymour, R., Gardner, A., Lambert, I., Fyson, R. & Wright, N. Journal of Public Health, 2024. Paper

How Can We Estimate Modern Slavery Globally?

Seymour, R. & Silverman, B. Chance, 36(4), 22-29, 2023. Paper