time series
Definition
Time series refers to sequences of repeated measurements collected from an individual over time, used to model the temporal dynamics of psychological variables such as emotions, cognition, and behavior. In idiographic psychological research, time series data are gathered through methods including Experience Sampling, Ambulatory Assessment, and Ecological Momentary Assessment, often spanning several hundred time points across periods exceeding four months. The fixed moderated time series model, a vector autoregressive based approach grounded in the state space framework, extends standard time series analysis by allowing all model parameters, including the innovation structure, to be moderated by contextual factors such as daily stress or social interaction. This makes it possible to estimate not only when and where a psychological network changes, but also which external variables drive that change.
Sources: Bringmann et al. (2024)
Related Terms
- networks (1 shared article)
- personalized models; moderation analysis (1 shared article)
- vector autoregressive model (1 shared article)
Applications
Time Series and Vector Autoregressive Models
Vector autoregressive models form the statistical backbone of idiographic time series analysis, capturing lagged relationships between psychological variables measured repeatedly within a single person. Advances in time series analysis have produced time-varying versions of these models, in which both mean levels and temporal dynamics are permitted to change across the observation period. The fixed moderated time series model extends this framework further by allowing moderation of all VAR parameters, including the innovation variance that reflects reactivity to unobserved factors.
Sources: Bringmann et al. (2024)
Time Series and Idiographic Psychological Networks
Idiographic psychological networks are estimated directly from intensive longitudinal time series data collected for a single individual, with the VAR model translating lagged associations among variables into network edges. Standard network models assume stationarity, meaning the connections and mean levels are held constant across the entire time series, an assumption that becomes untenable in longer observation windows where clinical change is expected. The fixed moderated time series model addresses this by revealing how network structure shifts as a function of contextual moderators, offering clinicians a more accurate representation of a patient's changing psychological system.
Sources: Bringmann et al. (2024)
Time Series and Moderation Analysis
Moderation analysis applied to time series data identifies which contextual factors, such as social interaction or sleep quality, are associated with changes in the parameters of a psychological network. Unlike time-varying models that locate when a change occurs without explaining it, the fixed moderated time series model tests whether a specified moderator accounts for shifts in lagged connections, mean levels, or innovation variance. Applying this approach to two patients with depression, the method demonstrated that both a binary moderator and a continuous moderator produced detectable changes across different parts of the emotion network.
Sources: Bringmann et al. (2024)
