Algorithmic divide among elderly and sandwich generation in Hong Kong: Development and validation an algorithm literacy scale

Algorithms are an integral part of our daily lives, shaping the selection and presentation of information and communications on the Internet. At the same time, media users face a lack of control and transparency when interacting with these systems, as algorithms remain largely a black box to end users. Furthermore, users' perceptions of the algorithm were found to be closely related to their perceived autonomy. When users feel they have control over their online interactions, they are less aware of the impact of the algorithms that govern their interactions (Dogruel et al., 2022).Previous research has developed conceptualization and dimensions of algorithmic literacy and provided specific scales for the components of each dimension. In addition, researchers have explored some potential relationships, such as the impact of demographic factors on algorithm literacy, and the impact of previous experience on algorithm cognition. There are some findings on how algorithm literacy helps users envision, understand, and use algorithms, depending on their understanding of the information flow control embedded in the algorithm (Shin et al., 2022).

Existing articles mainly focus on cognitive level of algorithm literacy, to understand how algorithms work, awareness and knowledge has been examined in the use of algorithms in online applications, platforms, and services. In the next stage, other dimensions should be explored to better connect users and algorithm mechanisms, such as critical evaluation and human-algorithm interaction. The discussion on the ability to critique algorithms and the skills to influence algorithms would enrich the conceptualization of algorithm literacy. To better understand the predictors on algorithm literacy and the components among dimensions, this study will discuss the difference in demographic characteristics and usage behavior that lead to variation in individuals’ algorithm literacy. Additionally, in the algorithm literacy component, to what extent is algorithmic awareness correlated with other dimensions. Based on an online survey in Hong Kong, the findings will provide initial evidence for the user pattern of algorithmic cognition and preliminarily explore the relationship between different dimensions of algorithmic literacy.

In this article, the authors firstly collected existing studies on algorithm literacy and reviewed sub-factors and measurement scales and analyzed previous studies to select the sub-factors and preliminary questions for the scale development. Then, confirmatory factor analyses were conducted to confirm the convergent validity and discriminant validity of the constructed scale. Linear regression analysis was conducted to confirm the predictive validity of the sub-factors of algorithm literacy. Based on the definitions and operations of existing research, this study developed algorithm literacy measurements. To test Hong Kong residents' perceptions of algorithmic platforms and the effectiveness of algorithm literacy scale, an online survey was conducted in October 2023 (N=788). Conducting confirmatory factor analysis, the results reveal that algorithm awareness, ethical awareness and privacy concern are crucial sub-factors of algorithm literacy in mobile applications usage. Meanwhile, using linear regression examines the impact of demographic factors and usage behaviors on algorithm literacy, as well as the relationship between different dimensions of algorithm literacy. The findings indicate that age and education are factors correlated to algorithm awareness. Furthermore, algorithmic awareness is positively correlated with ethical awareness and privacy concern. The conclusions provide evidence for the algorithm divide in Asia and enlighten the interaction between different dimensions in algorithm literacy.