A continuous-time approach to intensive longitudinal data: the way forward?”
Oisín Ryan: “A continuous-time approach to intensive longitudinal data: What, Why and How?” click here to access the slides
The increased availability of intensive longitudinal data has opened up new opportunities for researchers to investigate the dynamics of processes. The models that are used in this context can be broadly characterized as discrete-time (DT) or continuous-time (CT) models. While DT models are more popular, CT models promise to overcome conceptual and practical problems which have long been associated with DT approaches. This talk will provide a broad didactical treatment of the CT approach. Following the necessary background on DT and CT models, we will discuss the conceptual reasons why the CT perspective is valuable in moving our understanding of processes forward, and how these models can be interpreted once estimated from empirical data.
Sanne Booij: “Intra-individual cortisol dynamics in continuous time: An illustration”
I will present a study into daily life cortisol dynamics of two different samples with healthy and depressed individuals, as analyzed with a continuous time process model. The first sample was an all-female twin sample with experience sampling 10 times a day for 5 days. The second sample was a pair-matched sample with experience sampling three times a day for 30 days. The parameters of interest were set point (the point to which cortisol returns after any perturbation), variability around the set point, and speed of return (to the set point). The data were analyzed using a continuous-time process model, specifically a multilevel stochastic differential equation model (Oravecz et al. 2009; Oravecz et al. 2016). Despite the different characteristics of the samples and the very different sampling intervals, the model recovered strikingly similar parameters for the dynamics in both samples. The results will be discussed in light of the continuous time approach. If you want the slides, please send an e-mail to Sanne Booij. s.h.booij<at>umcg.nl