Using Genetic Algorithms in a Large Nationally Representative American Sample to Abbreviate the Multidimensional...

Area of Science:

• Psychometrics
• Machine Learning
• Psychological Measurement

Background:

• Experiential avoidance is crucial in psychological interventions, necessitating accurate measurement.
• The 62-item Multidimensional Experiential Avoidance Questionnaire (MEAQ) offers comprehensive assessment but is lengthy.
• Existing short forms may sacrifice multidimensionality, limiting their utility.

Purpose of the Study:

• To develop a shorter, multidimensional version of the MEAQ using genetic algorithms (GAs).
• To maintain the six-factor structure and psychometric integrity of the original MEAQ.
• To reduce participant burden while preserving measurement fidelity.

Main Methods:

• Utilized a GA-based scale abbreviation method (GAabbreviate R package).
• Applied the method to shorten the 62-item MEAQ to a 30-item version (MEAQ-30).
• Validated the MEAQ-30 in a large, nationally representative sample (N = 7884).

Main Results:

• The MEAQ-30 demonstrated near-identical performance to the 62-item MEAQ.
• Factor structure, correlations, and latent associations were preserved.
• The short form showed similar regional mean distributions as the original MEAQ.

Conclusions:

• The MEAQ-30 offers a valid, multidimensional measure of experiential avoidance.
• This GA-derived short form balances measurement accuracy with reduced participant burden.
• Machine learning methods show promise for advancing psychometric scale development.