Transitions and Resilience in Ecological Momentary Assessment: A Multiple Single-Case Study.

Abstract

Ecological momentary assessment (EMA) of affect, cognition and behavior aims to provide a 'window into a person's daily life'. But what should we look for through this window? In this paper, we compare a statistical perspective, grounded in probability theory, with a dynamic pattern perspective, grounded in complexity theory, on two common phenomena in EMA data: non-stationarity and outlying values. From a statistical perspective, these phenomena are considered nuisances that should be dealt with. From a dynamic pattern perspective, in contrast, non-stationarity may signal transitions from one dynamic pattern to another (e.g., a transition from a neutral to a persistent sad mood), whereas outlying values may signal recovery from perturbations (e.g., stressful life events). We evaluated the dynamic pattern view with a triangulation study of multiple single cases that took part in the Track your Mood EMA study, where participants reported on their emotions and daily events for 60 days. We found that non-stationarity was indeed related to a pattern transition, whereas outlying values were related to recovery after perturbations. These findings show that person-oriented EMA research would benefit from a dynamic pattern perspective that can identify highly meaningful and clinically relevant phenomena that are otherwise at risk of being missed. Complementing EMA time series with contextual information and qualitative data will be essential to genuinely understand these phenomena.