Understanding the Relationship between Acceptance of Multidimensional Health and Fitness Functions and Actual Usage of Wrist-Worn Wearables (Preprint)

Abstract

BACKGROUND The health and fitness functions of wrist-worn wearables (WWD), designed to encourage users to be physically active and take preventive measures against diseases, now provide increased health benefits due to their digitalized interactive environment. However, literature on the role played by the individual health and fitness functions of WWD (i.e., goal setting, data tracking, goal tracking, and share-archive function) in users’ physical activity is limited. OBJECTIVE The present study aims to identify existing users’ acceptance of the multidimensional health and fitness features of WWD (i.e., knowledge acquisition, perceived usefulness, and perceived ease of use) required for each stage of physical activity (i.e., before, during, and after) and examine the relationship between its acceptance and the actual use of its health and fitness attributes. METHODS Both qualitative and quantitative approaches were taken for analyzing the relationship. A focus group interview was conducted (n=9) to design the research model, including the operationalized definition of the constructs required for the study. A questionnaire survey was conducted with respondents in South Korea (n=480). Partial least squares structural equation modeling (PLS-SEM), via Smart PLS 3.0, was employed to test the hypotheses proposed in our study. RESULTS The participants were aged 10–70, and their careers in sport ranged from being national athletes to amateur- and student-athletes. The sample comprised 47.5% females and 52.5% males. For the measurement model analysis of lower-order constructs, all the model evaluation criteria (i.e., item loadings (ranging from .771 to .912 > .70), Cronbach’s alpha (ranging from .779 to .946 > .70), composite reliability (ranging from .859 to .955 > .70), convergent validity (AVE ranging from .653 to .780 > .50), and discriminant validity via Fornell-Larcker criterion and HTMT) met the requirement. Because we treated actual usage of WWD as a second-order construct, we tested validity and reliability of the higher-order constructs via the repeated indicators and two-stage approaches. As a result, the psychometric properties of second-order constructs were found to be adequate. With respect to the predictive relevance of the hypothesized structural model, the R2 (.725, t=31.366, P <.001) and the Q2 value (.469) indicated a high predictive accuracy of our hypothesized model. Further analysis revealed that all the paths were significant: from goal setting to actual usage (β=.133, t=1.756, P=.001), goal tracking to actual usage (β=.281, t=5.158, P<.001), and share-archive to actual usage (β=.400, t=8.399, P<.001) except the path from data tracking to actual usage. CONCLUSIONS The results showed that when users learn to use fitness functions and perceive them as useful for physical activity without causing any difficulty, they tend to use those functions more, which provides enhanced health benefits under the digitalized interactive environment of WWD.