Estimating Latent State-Trait Models for Experience-Sampling Data in R with the lsttheory Package: A Tutorial

Area of Science:

• Psychological Methods
• Quantitative Psychology
• Personality Psychology

Background:

• Experience-sampling methodology (ESM) is increasingly used, necessitating advanced analytical techniques.
• Distinguishing transient situational effects from stable individual differences is crucial in ESM.
• Latent state-trait (LST) models offer a framework for this differentiation.

Purpose of the Study:

• To provide a tutorial on multiple-indicator wide-format LST models for ESM data.
• To introduce user-friendly software (browser app and R-function in 'lsttheory') for specifying LST models.
• To demonstrate the application of LST models in analyzing well-being dynamics.

Main Methods:

• Discussion of first-order and second-order LST model specifications and their assumptions.
• Introduction of a new R-package 'lsttheory' with a browser app for LST model specification.
• Application of LST models to a five-day experience-sampling study on well-being.

Main Results:

• An autoregressive model with indicator-specific traits was optimal for the ESM data.
• Results indicated high consistency in well-being, suggesting a stronger influence of person-level traits than situational factors.
• Extraversion, emotional stability, and agreeableness predicted trait well-being.

Conclusions:

• LST models, particularly with the introduced software, offer a flexible yet accessible approach for ESM data analysis.
• Well-being appears to be predominantly influenced by stable personality characteristics.
• Recommendations for model fit assessment and comparative analyses are provided.