A Multi-Rater Latent Growth Curve Model.

Area of Science:

• Psychometrics
• Longitudinal Data Analysis
• Statistical Modeling

Background:

• Multiple raters are frequently used to assess psychological and social-emotional constructs to mitigate subjective bias.
• Longitudinal data with multiple raters are increasingly common, yet models for estimating growth in such data are scarce.

Purpose of the Study:

• To develop and validate a statistical model for estimating growth from longitudinal data collected by multiple raters.
• To address the challenge of unique rater variance in multi-rater longitudinal datasets.

Main Methods:

• A novel latent growth curve model was developed to separate shared rater perceptions from unique rater variance at each time point.
• The model integrates shared rater variance with a latent growth curve framework to analyze changes over time.
• Model performance was evaluated using simulation studies and empirical data analysis.

Main Results:

• The proposed model effectively recovers true growth parameters in longitudinal multi-rater data.
• The model demonstrates superior performance compared to simpler methods, such as growth estimation based on a single rater.
• The approach successfully removes unique rater variance, isolating shared perceptions for growth analysis.

Conclusions:

• The developed model offers a robust method for estimating growth from longitudinal data involving multiple raters.
• This approach provides researchers with a valuable tool for analyzing complex multi-rater longitudinal datasets, enhancing the accuracy of growth parameter estimation.