Twins and Causal Inference: Leveraging Nature's Experiment.

Area of Science:

• Behavioral Genetics
• Biostatistics
• Epidemiology

Background:

• Twin studies are crucial for disentangling genetic, shared environmental, and nonshared environmental influences on population variance.
• Understanding these influences is fundamental for biometric decomposition and causal inference in research.

Purpose of the Study:

• To review how twin data strengthens causal inference by controlling for shared influences on exposures and outcomes.
• To explore analytical approaches for assessing if exposure-outcome associations persist after accounting for shared etiology.
• To highlight limitations and considerations for using twin data in causal hypothesis testing.

Main Methods:

• Biometric decomposition of population variance using twin data.
• Multivariate structural equation models.
• Cotwin control methods, direction of causation models (cross-sectional and longitudinal), and extended family designs.

Main Results:

• Twin data allows for controlling shared genetic and environmental factors, strengthening causal inference.
• Various analytical methods can assess the robustness of exposure-outcome associations against shared etiology.
• The review synthesizes approaches for interrogating causal hypotheses using twin and family designs.

Conclusions:

• Twin studies offer a powerful framework for causal inference by isolating genetic and environmental effects.
• Researchers must consider specific limitations when employing twin data to test causal hypotheses.
• The discussed methods provide a comprehensive toolkit for advancing etiological research.