Stress processes have long been of interest to researchers. A growing body of research explores the bidirectional relations between stressors, psychological and physical states. However, research on stress processes and their individual differences from a dynamic systems perspective is still lacking. This study examined dynamic feedback loops between stressors, psychological distress, and physical symptoms simultaneously using ecological momentary assessment. Three hundred and fifty-six participants completed five momentary assessments on stressors, psychological distress, and physical symptoms per day for 7 days in 2023. They also completed measures of their Big Five personality traits, depressive symptoms, and anxiety symptoms. Dynamic structural equation models showed positive cross-lagged and feedback effects of stressors with psychological distress and physical symptoms, suggesting their self-perpetuating loops. Agreeableness and conscientiousness were protective factors, and neuroticism was a risk factor for the stressor-psychological loop. Individuals with depressive and/or anxiety symptoms were characterized by (a) greater inertia of psychological distress and physical symptoms, (b) stronger reverse effects of psychological distress and physical symptoms on subsequent stressors, (c) significant reciprocal effects between psychological distress and physical symptoms, and (d) stronger self-perpetuating loops of stressors with psychological distress and physical symptoms. Our findings highlight the importance of analyzing feedback loops to understand bidirectional relations and individual differences in dynamic stress processes, providing insights for relevant personalized interventions. (PsycInfo Database Record (c) 2024 APA, all rights reserved).
6 July 2023 in Structural Equation Modeling: A Multidisciplinary Journal
Differential equation models have become increasingly popular for investigating dynamic processes. However, commonly used two-stage estimation methods, such as the generalized local linear approximation (GLLA), often produce biased parameter estimates. This study proposes a bias-correction method for GLLA estimates in second-order differential equation models. The method solves a bias-correction equation (derived from the relation between true parameter values and their biased estimates) via stochastic approximation, producing asymptotically unbiased estimates even with large initial bias. We first demonstrate the application of the bias-correction method by correcting the bias of a single parameter in a simple second-order differential equation model (i.e., the linear oscillator model with time-independent measurement error). We then extend the method to a more commonly used second-order differential equation model (i.e., the damped linear oscillator model), examining its performance in simultaneously addressing multiple parameters and incorporating time-dependent dynamic error. A simulation study shows that the bias-correction method substantially reduces bias in GLLA estimates, yielding highly accurate and precise parameter estimates. An empirical illustration further compares the results of GLLA and the bias-correction method. Our findings highlight the effectiveness of the proposed method in improving parameter estimation for differential equation models, offering an enhanced approach for analyzing dynamic processes.
Autoregressive (AR) models are widely used to study within-person dynamics in intensive longitudinal data (ILD), but they assume constant parameters, an assumption that may not hold when the dynamics change, as during development, intervention, or recovery. The time-varying autoregressive (TV-AR) model accommodates such change by representing the intercept and AR coefficient as smooth functions of time. As originally proposed, however, it was developed for a single individual and evaluated with time-averaged accuracy metrics. We address these limitations by extending it to a Bayesian multilevel formulation with population-average and individual-specific time-varying parameters, and by emphasizing pointwise evaluation of accuracy rather than aggregate accuracy alone. Two simulation studies examined how well the single-level and multilevel models recover time-varying dynamics. Examining bias at each time point, we found that it was largest near the series boundaries and in regions of steep change or high curvature. The time-varying intercept and AR coefficient showed complementary biases that largely cancelled in the implied mean, which was therefore recovered with smaller bias. In the multilevel model, the best allocation between persons and time points depended on the parameter and inferential level: the AR coefficient favored longer series at both levels, whereas the implied mean favored more persons for population-average recovery but longer series for individual recovery. We discuss implications and recommendations for study design, result interpretation, and model selection.
20 January 2026 in Applied Psychology: Health and Well-Being