Papers for

epidemiologists

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Backtracking methods offer new ways to measure direct and indirect effects

Backtracking Direct, Indirect and Total Effects

Abstract: Natural direct and indirect effects are defined through cross-world interventional counterfactuals. We propose an alternative construction based on backtracking counterfactuals. Rather than intervening on the treatment or mediator, we use a backtracking conditional distribution to construct an alternative exogenous state satisfying a desired treatment-mediator constraint. This yields the backtracking direct and indirect effects, and also the backtracking total effect. Likewise, we adapt the population intervention effects.

Tue 20 OctMachine Learning
The gist
When studying cause and effect, it’s useful to separate direct influences from those that happen indirectly. The existing way to do this uses ‘cross-world’ ideas that imagine what would happen under different interventions, which can be complicated. This paper presents a different approach called 'backtracking,' which changes how the cause and intermediate factors relate without actively intervening. This new method creates alternative measures of direct and indirect effects that could improve how we understand causal relationships.
Open → 2610.28910v1