Examination of the Problem Gambling Severity Index for Use with Older Adults: A Rasch Model Approach

Area of Science:

• Gerontology
• Psychometrics
• Addiction Research

Background:

• The nine-item Problem Gambling Severity Index (PGSI) is commonly used to assess problem gambling severity in older adults.
• Evaluating the PGSI's appropriateness for this demographic is crucial for accurate score interpretation.
• Previous research has not utilized Rasch modeling to examine the PGSI's psychometric properties in older adults.

Purpose of the Study:

• To evaluate the psychometric properties of the PGSI in older adults using Rasch modeling.
• To assess the dimensionality, item fit, response utility, and differential item functioning (DIF) of the PGSI in this population.
• To identify potential modifications to improve the PGSI for older adult gamblers.

Main Methods:

• A Rasch rating scale model was applied to data from the Quinte Longitudinal Study (n=571, aged 60+).
• Analysis assessed PGSI item dimensionality, fit, response option utility, and DIF by gender.
• A larger sample (N=3206, aged 18-80+) was used to examine DIF across age subgroups.

Main Results:

• The PGSI demonstrated unidimensionality and acceptable item polarity.
• Three items exhibited model misfit, and response options were underutilized.
• Model-person separation and reliability were poor, while model-item separation and reliability were excellent.
• No significant gender-based DIF was found in older adults, but three items showed DIF between older and younger adults.

Conclusions:

• The PGSI shows promise for assessing problem gambling in older adults, but requires modifications for improved measurement.
• Specific items may function differently across age groups, necessitating further investigation and potential revision.
• The study highlights the utility of Rasch modeling in evaluating measurement tools for specific demographic groups.