Stochastic counterfactuals and stochastic sufficient causes

Area of Science:

• Causal Inference
• Statistical Modeling
• Epidemiology

Background:

• Causal inference primarily focuses on deterministic counterfactuals.
• Stochastic counterfactuals, where outcomes vary probabilistically, are less explored.
• Deterministic sufficient cause frameworks model counterfactuals using causal mechanisms.

Purpose of the Study:

• To extend the deterministic sufficient cause framework to stochastic counterfactuals.
• To formally define a stochastic sufficient cause framework.
• To investigate the conditions for detecting sufficient cause interactions in stochastic settings.

Main Methods:

• Development of formal definitions for stochastic sufficient causes.
• Extension of deterministic sufficient cause interaction testing to the stochastic domain.
• Application of the framework to genetic examples.

Main Results:

• Formal definitions for a stochastic sufficient cause framework are provided.
• Empirical conditions for detecting sufficient cause interactions are shown to be consistent between deterministic and stochastic settings.
• The framework is illustrated using examples from genetics.

Conclusions:

• The proposed stochastic sufficient cause framework is a valid extension of deterministic models.
• Detecting sufficient cause interactions in stochastic settings can utilize established empirical conditions.
• This work enhances causal inference methodologies for probabilistic outcomes, particularly in fields like genetics.