Theoretical note: The relation between structure and dynamics in psychological networks of attitudes

Two claims of the the Causal Attitude Network (CAN) model and the descendent Attitude Entropy framework (AE) are indicative of significant theoretical hurdles facing the psychological network modeling efforts of attitudes. The first claim is that the dynamics of change in an Ising-like attitude network, under perturbation of any one single node, can be inferred from the static network attributes of said node. The second claim is that psychological network models of attitudes with Ising-like dynamics will maximize both attitudinal consistency and accuracy when within the small-world topological regime. The first claim, one with significant application potentials, has not been sufficiently tested; the second claim, one with high theoretical novelty, has never been addressed. Using a set of analytic results and simulations, we found little support for these claims–in short, the predictions are not logically consistent with the theory. Our results have implications beyond attitude models to the the larger field of psychological networks (e.g., in clinical psychology) in reference to how we should explain and understand their dynamics. KEYWORDS: attitudes, neural networks, dynamical systems, psychological networks.