Estimating the Joint Probability Density for Index Construction: Some Simplifications Using the TWEAK as Example

Area of Science:

• Psychometrics
• Quantitative Psychology
• Statistical Modeling

Background:

• The joint probability density (JPD) method offers robust index construction with minimal data restrictions.
• Estimating JPD can be challenging with a large number of input variables.

Purpose of the Study:

• To evaluate two simplified JPD estimation methods for the TWEAK problematic drinking screener.
• To compare these simplified JPD methods against fully estimated JPD, conventional TWEAK scoring, confirmatory factor analysis (CFA), and item response theory (IRT).

Main Methods:

• A mailed survey was administered to 410 gender-stratified, post-9/11 Veterans with pending PTSD disability claims.
• The study assessed problematic drinking using the TWEAK screener and various scoring methodologies.

Main Results:

• Simplified JPD estimations showed high concordance (Spearman's correlations 0.97-1.00) with fully estimated JPD.
• All JPD methods identified gender differences in TWEAK scores related to alcohol use diagnosis, though not consistently for women.
• IRT analysis revealed local dependence issues, while JPD methods demonstrated high agreement with CFA and conventional scoring.

Conclusions:

• Simplified JPD estimations are computationally easier and yield results comparable to the full JPD.
• All three JPD estimation approaches demonstrated strong concordance with traditional TWEAK scoring and CFA methods.