Can cross-lagged correlations prove causation in social science? This book asks hard questions and tests them with clear examples.
This work explains how choosing different time periods can change the causal directions suggested by cross-lagged analysis. It also shows that non-linear feedback between variables can make these methods unreliable. Through a simple simulation, the author demonstrates how two variables in a social system can appear to drive each other, switch roles, or seem unconnected depending on how data are observed. The goal is to identify when the cross-lagged method is valid and when it is misleading, so researchers can design better studies.
- Understand how time window choices affect causal inferences from cross-lagged data.
- See why non-linear feedback can disguise true relationships between variables.
- Explore a simulation that highlights competing interpretations of causality.
- Learn practical tests and diagnostics to determine when cross-lag analysis applies.
Ideal for readers interested in behavioral science methods, research design, and statistical inference who want a cautious, evidence-based view of this approach.