Psychometric evaluation and item response theory analysis of the COVID Stress Scales in an older adult population.

Abstract

Objectives: The COVID Stress Scales (CSS) represent a widely used self-report measure of stress and anxiety-related responses to COVID-19. Although the CSS have been validated across various nations and languages, their psychometric properties have not been assessed at the factor- or item-level with older adults. We aimed to psychometrically evaluate the CSS in older adults.

Method: The CSS was examined with 486 North American older adults aged 65 years and older. Data were collected in January 2024 using Qualtrics Panels. Reliability was assessed using Cronbach's alpha and McDonald's omega coefficients, structural validity using confirmatory factor analysis, and discriminant validity using a social desirability measure. Item properties were examined using item response theory.

Results: The CSS demonstrated robust internal consistency and a defensible five- and six-factor structure, with the six-factor providing the most optimal model of measurement. All items adequately discriminated among respondents with varying levels of COVID-related stress.

Conclusion: This study is first to demonstrate that the CSS functions as a reliable and valid tool for evaluating COVID-related stress among older adults, a necessary step for supporting its use in assessing mental health impacts of pandemics in a population at high risk of negative post-infection outcomes.