Objective: Suicide risk is a nonlinear temporal process, but the ways in which suicide-focused interventions have statistically examined risk effects have ignored these nonlinearities. This paper highlights the potential benefits of using data analytic methods that account for nonlinear change patterns.
Method: Using a dynamical systems perspective, interventions are framed in terms of attractor dynamics. An attractor has three primary qualities where an intervention can have an effect. These correspond to contextual differences, shifts in the underlying temporal patterns, and changes in the stability of the temporal pattern.
Results/conclusions: It is argued that the ideal effect is one in which there is both an observed change in stability and a shift in the underlying temporal pattern toward less risk. Other types of intervention effects can have alternate explanations that are less desirable. Mean, variance, and growth differences are discussed within a systems context, and an example model is provided using Latent Change Score Modeling (McArdle, Annual Review of Psychology, 60, 2009, 577-605).
© 2020 The American Association of Suicidology.