Post

The idea: Turn the target review into a living causal model.The model maps out what has to be true for a program to work, with each link tied back to the evidence behind it, so a researcher can interrogate any step. The output is a go / no-go decision rather than a summary.