A Moderated Nonlinear Factor Model for the Development of Commensurate Measures in Integrative Data Analysis

Area of Science:

• Psychometrics
• Quantitative Psychology
• Data Science

Background:

• Integrative Data Analysis (IDA) pools data from multiple sources to increase statistical power and heterogeneity.
• A key challenge in IDA is creating valid psychometric scores from data with varying items and response options.
• Previous work proposed moderated nonlinear factor analysis (MNLFA) for scoring within IDA.

Purpose of the Study:

• To develop a general framework for estimating MNLFA models and obtaining scale scores across diverse settings.
• To extend previous MNLFA methods for enhanced application in Integrative Data Analysis.
• To examine the factor structure of depressive symptomatology using a large, multi-study dataset.

Main Methods:

• Proposed a five-step procedure for estimating MNLFA models within an IDA framework.
• Applied the procedure to data from 1972 individuals (ages 11-34) pooled across three independent studies.
• Utilized 17 binary items assessing depressive symptomatology to test the factor structure.

Main Results:

• Successfully estimated MNLFA models and obtained individual-specific scale scores.
• Provided substantive conclusions regarding the factor structure of depression.
• Demonstrated the utility of the proposed framework for handling complex, multi-source data.

Conclusions:

• The developed MNLFA framework offers a robust method for scoring in Integrative Data Analysis.
• The findings contribute to understanding the factor structure of depression across a wide age range.
• Recommendations are provided for practical implementation of these advanced psychometric methods.