Reproducibility of ecological data

After twenty years of procrastination one of our bluest datasets has finally been analysed! The data was one of two sets of data used in a ‘many analysts’ study to look at how the choices that are made in an analysis pathway affects outcomes. The study is covered in a news story published today in Nature.

https://www.nature.com/articles/d41586-023-03177-1

The dataset was one of two that were analysed by 246 biologists in the study which is available online. The variation in the outcomes was remarkable given that all analysts started with the same dataset and were addressing the same question – how is nestling growth affected by the number of siblings that offspring have in the nest. There is continuous variation in the effect that analysts found. A number of analyses found significant positive effects, many found no effect and the majority found significant negative effects of siblings on growth. The study is revealing about the robustness of results in ecology and should hopefully help lead to a different approach to analyses in the future.

Leave a comment