Results of d-separation test.

For each model, the basis set lists the statistically testable predictions of independence made by each model. Also given is Fisher’s C statistic and the associated p-value for each path model. Non-significant C values suggest that we can accept the proposed model. K is the total number of free parameters in a model. AICc is Akaike’s information criterion of model selection and ∶AICc is the relative difference in AICc between a given model and the best-fitting model (b-Generalization).