Modeling categorical time-to-event data: The example of social interaction dynamics captured with event-contingent...

Area of Science:

• Psychology
• Data Science
• Behavioral Science

Background:

• Ambulatory assessment collects vast data on daily behaviors using active and passive methods.
• Understanding the timing and types of social interactions is crucial for psychological research.
• Traditional methods have limitations in analyzing complex event data from daily life.

Purpose of the Study:

• To illustrate the application of multilevel and multistate survival analysis in ambulatory assessment research.
• To model the dynamics of social interactions, including their timing and categories.
• To provide a tutorial for analyzing intensive longitudinal data using survival models in R.

Main Methods:

• Utilized multilevel and multistate survival analysis techniques.
• Applied survival models to intensive longitudinal data, specifically event-contingent reports.
• Demonstrated modeling of social interaction timing and categories using the R statistical programming language.

Main Results:

• Survival analysis effectively models the timing and categories of social interactions captured in daily life.
• The study provides a practical framework for analyzing complex interpersonal dynamics.
• Empirical application with 64,112 events from 150 participants validated the approach.

Conclusions:

• Survival analysis offers a powerful approach to advance the understanding of social interaction dynamics in ambulatory assessment.
• This methodology can uncover novel insights into behavior patterns in naturalistic settings.
• The integration of survival models enriches the analytical toolkit for behavioral researchers.