JMIR Publications

ABSTRACT

Background:

Global population ageing makes digital technologies increasingly important for older adults to maintain function and participation in meaningful activities. Esports, an emerging form of digital sports and leisure, is beginning to cater to older population. However, existing instruments measuring esports participation intention and attitudes were developed for adolescent or young adult samples, and little psychometric evidence exists to indicate whether their normative referents and control-belief content are transferrable to gerontechnical digital leisure contexts.

Objective:

To examine the psychometric properties of an age-adapted Theory of Planned Behavior (TPB)-based esports participation intention questionnaire (TPB-Esport-Q) among community-dwelling older adults in Hong Kong, including the prespecified nine-factor structure, reliability, convergent validity, discriminant validity, and measurement invariance across gender.

Methods:

Building on the original questionnaire, the adaptation for older adults used qualitative interview findings on esports participation among middle-aged and older adults in Hong Kong. Psychometric testing was conducted in a mixed-recruitment sample of 802 older adults (496 women; 353 esports participants recruited on-site at community centers, and 449 older adults with no esports participation in the past 6 months recruited through the random telephone survey). Confirmatory factor analysis used robust maximum likelihood estimation with full-information maximum likelihood for missing data. Internal consistency was evaluated using Cronbach α, McDonald ω, and composite reliability; convergent validity using average variance extracted; and discriminant validity using the heterotrait–monotrait ratio (HTMT). Measurement invariance across gender was tested with multi-group confirmatory factor analysis, and measurement equivalence across data sources alongside a retrospective stratified random split-half internal cross-validation served as sensitivity analyses.

Results:

The final 37-item nine-factor model retained only the ATT2–ATT3 within-construct residual correlation, and the overall model fitted well (χ² (592) = 1433.21, CFI=0.945, TLI=0.939, RMSEA=0.048, SRMR=0.051); composite reliability was 0.620–0.948, average variance extracted 0.365–0.858, and the maximum HTMT 0.849. Measurement invariance across gender was approximately supported up to the strict level. In the retrospective internal cross-validation, fit was similar in the development (n=400) and validation (n=402) samples (CFI=0.943 vs 0.941; TLI=0.936 vs 0.933; RMSEA=0.050 vs 0.051; SRMR=0.057 vs 0.054). The cross-data-source sensitivity analysis supported configural, approximate metric, and partial scalar invariance, but not full scalar invariance.

Conclusions:

The TPB-Esport-Q showed an acceptable nine-factor structure in this Hong Kong community-dwelling older adult sample, and discriminant validity and measurement invariance across gender were generally supported. Participation intention, attitude, subjective norm, and most belief constructs performed relatively stable, whereas the reliability and convergent validity of perceived behavioral control and control beliefs–barriers were relatively weak. Internal cross-validation reduced concerns about model optimization in a single sample, but external validation in a fully independent sample and further refinement of the control-related items remain necessary. Clinical Trial: This study adhered to the ethical principles of the Declaration of Helsinki, and the study protocol was reviewed and approved by the Human Research Ethics Committee of The Education University of Hong Kong (approval number: 2021-2022-0107).