R code for WHO SEJ study 2026

R code used to analyze experts' source attribution assessments and aggregate them by using the Classical Model for Structured Expert Judgement.

The WHO SEJ study aimed to quantify uncertainty regarding source attribution for 40 hazards across 17 subregions. Expert assessments for 6 major pathways and 14 food groups were collected via Qualtrics. Expert uncertainty assessments were validating using two objective scores (calibration and information score) and weights were derived to aggregate their distributions. Aggregated distributions (DMs) were also validated using the same objective scores and the best performing DM was chosen. 10,000 observations were collected from these distributions, and the samples were normalized, such that mean estimates of proportion of major pathways sum to 100% and mean estimates of food groups sum to 100%.