CONSTRUCTION AND VALIDATION OF ALGORITHM ANXIETY SCALE

Abstract

Although algorithm-driven digital environments are increasingly prevalent, they introduce emerging psychological challenges, and validated instruments for measuring algorithm-related anxiety remain limited. The purpose of this study was to create and validate the Algorithm Anxiety Scale (AAS), a multidimensional scale of emotional and cognitive reactions to algorithmic systems. A sequential mixed-method design was used. Phase 1 involved qualitative interviews with social media users to gain insights into their lived experiences of being exposed to algorithm-driven content, which informed the development of items based on both empirical themes and theory such as techno-stress theory, surveillance capitalism and uncertainty management theory. The refined scale was quantitatively validated with 254 participants (18 to 60 years) during Phase 2. The two factors, Emotional Algorithm Anxiety (EAA) and Algorithmic Control and Uncertainty Perception (ACUP), were found to be stable through exploratory factor analysis. The model demonstrated acceptable fit based on confirmatory factor analysis indices (CFI = .961, RMSEA = .061, and SRMR = .066). The results of reliability analysis showed that the overall scale had a high internal consistency of α = .902, EAA of α = .871, and ACUP of α = .781. Construct validity was determined by composite reliability and inter-factor correlations. The findings suggest that algorithm anxiety represents a measurable psychological construct with affective and cognitive components related to algorithmic systems. The AAS offers a legitimate and reliable tool for further studies on digital behavior, algorithmic influence, and psychological well-being in technology-mediated environments.