Improved estimation of autoregressive models through contextual impulses and robust modeling

Area of Science:

• Psychology
• Affective Science
• Longitudinal Data Analysis

Background:

• Dynamic paradigm in affect research aims to characterize daily affective processes.
• Contextual conditions and events significantly influence affective processes.
• Explicitly modeling context faces challenges when events are hard to define or measure.

Purpose of the Study:

• To investigate the impact of contextual events on affective processes.
• To examine how short-lasting contextual events affect emotions.
• To demonstrate the advantages of robust autoregressive models in handling contextual contamination.

Main Methods:

• Utilizing intense longitudinal data.
• Employing dynamic, autoregressive-type models.
• Focusing on events with short-lasting main effects on affective processes.

Main Results:

• Contextual events can act as hidden confounders obscuring affective dynamics.
• Specific forms of contamination can beneficially trigger and leverage autoregressive dynamics.
• Robust autoregressive models outperform classical models in managing contextual contamination.

Conclusions:

• Robust autoregressive models are effective in characterizing daily affective processes.
• These models capitalize on positive leverage effects from contextual events.
• They mitigate negative obscuring effects of contextual events on emotional dynamics.