Automated Inductive Causation: Recovering Causal Structures from Observational Data
Abstract: Electronic diaries continuously generate information-rich observational data. The underlying causal network from such data could provide researchers and therapists with valuable information. The inductive (inferred) causation algorithm (IC algorithm) is a promising technique for recovering causal structures from observational data, yet has never been formally examined for its use for psychological data containing measurement error. This study examines how well, and under what circumstances, the algorithm accurately recovers causal structures from observational data. Four causal networks are created, embedded in different datasets, and then recovered using an automated IC algorithm. The results may lead to a better understanding of the characteristics of the IC algorithm, when applied to psychological data. Moreover, it could provide researchers and therapists with a tool for getting valuable causal information from observational data.
Included: Bachelor's thesis and scripts.