Sibling Models Can Test Causal Claims without Experiments: Applications for Psychology.

Area of Science:

• Psychological research
• Behavioral genetics
• Epidemiology

Background:

• Randomized experiments are ideal for causation but often impractical or unethical.
• Quasi-experimental designs use covariates to control confounds, but may miss gene-environment interactions.
• Existing methods can introduce bias due to unaddressed familial confounding.

Purpose of the Study:

• Advocate for genetically informed designs to strengthen causal inference.
• Introduce an analytic method based on the reciprocal standard dyad model for sibling comparisons.
• Demonstrate the efficacy of controlling for familial confounding in psychological research.

Main Methods:

• Adapted the reciprocal standard dyad model for kin pair analysis.
• Utilized discordant-kinship models to control for genetic and environmental confounds.
• Compared covariate-based approaches with genetically informed methods using four vignettes.

Main Results:

• Covariate-based analyses found significant associations in all four vignettes.
• Genetically informed methods revealed divergent results, highlighting familial confounding.
• The discordant-kinship model effectively controlled for background gene-and-environmental variance.

Conclusions:

• Familial influences are critical and often unaddressed in psychological research.
• Genetically informed designs offer a powerful tool for robust causal inference.
• The adapted dyad model provides a versatile and effective method for analyzing familial data.