Understanding Intelligent Personal Assistant Acceptance Across Generations in Behavioral Technology Contexts: A Multi-Group Serial Mediation Model
Journal of Technology in Behavioral Science, 2026 (Scopus)
- Yayın Türü: Makale / Tam Makale
- Basım Tarihi: 2026
- Doi Numarası: 10.1007/s41347-026-00729-w
- Dergi Adı: Journal of Technology in Behavioral Science
- Derginin Tarandığı İndeksler: Scopus, Psycinfo
- Anahtar Kelimeler: Anthropomorphism, Behavioral technology, Generational differences, Human-computer interaction, Intelligent personal assistants
- İstanbul Üniversitesi Adresli: Evet
Özet
Intelligent personal assistants are increasingly embedded in behavioral technology settings, yet generational differences in acceptance mechanisms are usually inferred from descriptive comparisons. This study tests how perceived competence relates to intention to use IPAs across Digital Native and Digital Immigrant cohorts within an integrated TAM–HCI model. Survey data from 585 respondents (422 Digital Natives, 163 Digital Immigrants) were analyzed using item-level CFA, measurement invariance testing, multi-group SEM with formal cross-group equality constraints, and bootstrapped indirect effects (5,000 resamples). Perceived anthropomorphism and perceived usefulness were modeled as mediators between competence and intention. In both cohorts, competence predicted anthropomorphism and usefulness, usefulness was the strongest proximal predictor of intention, and the serial competence→anthropomorphism→usefulness→intention pathway was significant. Of six structural paths, only competence→anthropomorphism differed across cohorts, with a stronger coefficient among Digital Immigrants; the difference persisted after adjustment for usage frequency and self-reported knowledge. Anthropomorphism forms part of the structural acceptance mechanism, whereas cohort membership qualifies one link rather than reorganizing the model. Cross-sectional convenience sampling, unequal cohort sizes, and limited direct IPA experience constrain inference and generalizability. Generational claims in acceptance research should be tested with explicit parameter constraints; future studies should incorporate validated literacy measures and behavioral outcomes.