Distinguishing cause from effect in psychological research: An independence-based approach under linear non-Gaussian...

Area of Science:

• Psychology
• Causal Inference
• Statistics

Background:

• Causal direction identification is crucial in psychological research.
• Observational data presents challenges in determining cause and effect.
• Existing methods struggle with hidden confounders.

Purpose of the Study:

• To introduce an independence-based approach for causal discovery between two variables.
• To develop a two-step algorithm for determining causal direction under linear non-Gaussian models.
• To address the challenge of hidden confounders in psychological studies.

Main Methods:

• An independence-based approach utilizing distance correlations.
• A two-step algorithm for empirical causal discovery.
• Monte-Carlo simulations for performance evaluation.

Main Results:

• The proposed algorithm effectively detects causal direction between variables.
• The method performs well even with weak hidden confounders.
• Distance correlations offer insights into the magnitude of confounding.

Conclusions:

• The developed algorithm provides a robust method for causal discovery in psychology.
• The approach is effective under realistic conditions with hidden confounders.
• This method enhances the ability to infer causality from observational psychological data.