Miguel A. Hernán, John Hsu and Brian Healy (2019). CHANCE.

https://doi.org/10.1080/09332480.2019.1579578

Author manuscript: https://arxiv.org/pdf/1804.10846

The authors distinguish describing data, predicting unknown outcomes and answering causal questions. They explain why predictive accuracy alone does not establish which intervention would help. This source supplies conceptual distinctions for an original fictional example in the graph; its visit counts are not from the paper.