The main aim of Stress in Action (SiA) is to “(…) gain insight into the causes and consequences of daily life stress, and to provide a path towards more stress-resilient citizens.” This paper, by SiA researchers dr. Jeroen Mulder and prof. dr. Ellen Hamaker, compares two methods for studying such causal mechanisms: A cross-lagged panel modeling approach within the structural equation modeling framework, popular in the social sciences, versus the use of structural nested mean models (SNMMs) with G-estimation, which is more common in biomedical disciplines. The latter approach is promoted in the causal inference literature as being a more robust methodology, and it is worthwhile for researchers to incorporate this method in their methodological toolbox. However, the causal inference literature can be difficult to get into due to its sometimes technical nature, and a disconnect with the modeling practices that researchers might be familiar with.
Jeroen Mulder: “In this paper, we bring the literature on linear SNMMs and G-estimation closer to those researchers. To aid this comparison, we use an substantive example from psychological practice throughout. We hope that this paper clarifies the modeling choices that (SiA) researchers have, and that it allows for better-informed decision making about which particular method is most useful for a researcher’s purpose.”
Joint Effects in Cross-Lagged Panel Research Using Structural Nested Mean Models. Mulder, J.D., Usami, S., & Hamaker, E.L. Structural Equation Modeling: A Multidisciplinary Journal, 2024.
