Abstract:

[Purpose/Significance] A thorough understanding of algorithm aversion among social media users, encompassing its manifestations and underlying causes, is crucial in the algorithmic era. This understanding serves as the cornerstone for accurately capturing users' information needs and preferences, which are constantly evolving due to technological advances and changes in societal behaviors. By studying how users perceive, interact with, and respond to algorithmic recommendations and personalizations, researchers can gain insight into the effectiveness and limitations of current algorithmic technologies. These insights are invaluable for improving and optimizing algorithms to ensure that they not only meet user expectations, but also enhance their overall experience and satisfaction. Moreover, understanding algorithm aversion can help design more ethical and transparent algorithms, foster trust between users and technology, and ultimately promote the sustainable development of the digital economy. In addition, this research has broader implications for the fields of human-computer interaction, artificial intelligence, and social media studies. By exploring the psychological, social, and cultural factors that influence users' attitudes and behaviors towards algorithms, researchers can contribute to the development of more user-centered and socially responsible technologies. This, in turn, can lead to more inclusive and equitable digital environments, where everyone can benefit from the advances of technology. [Method/Process] This study employed a qualitative research approach, which is well suited for exploring complex and nuanced phenomena such as algorithm aversion among social media users. Qualitative research allows for the collection of rich, detailed, and contextually embedded data, enabling a deeper understanding of the subject matter. To accomplish this, the study included in-depth interviews with 26 respondents, who were selected for their active use of social media and their diverse experiences and perspectives on algorithmic recommendations and personalizations. The interviews were conducted using a semi-structured format that allowed for flexibility in the conversation while still addressing key research questions and themes. This approach allowed the researchers to gain detailed insights into the participants' attitudes, beliefs, and experiences with algorithms, as well as their perceptions of the consequences of algorithm aversion. A rigorous coding process was used to analyze the collected data. This involved breaking down the textual data into smaller, manageable units, or codes, which were then categorized and grouped based on common themes and patterns. The coding analysis focused on three main areas: the expression of algorithm aversion, the formation mechanisms of algorithm aversion, and the consequences of algorithm aversion for social media users. Drawing on qualitative research paradigms, the analysis resulted in the construction of a theoretical model analysis framework specifically tailored to algorithm aversion among social media users. This framework provides a structured way to understand the complex interplay between users' attitudes, beliefs, and behaviors towards algorithms, and the factors that influence these attitudes and behaviors. The framework also highlights key consequences of algorithm aversion, such as reduced trust in social media platforms, decreased engagement with algorithmic recommendations, and potential negative impacts on user experience and satisfaction. [Results/ [Conclusions] The results reveal three distinct forms of algorithm aversion among social media users: algorithmic interruption, algorithmic complaint, and algorithmic evasion. These forms have significant implications for individuals, organizations, and society. Additionally, the study identifies personal factors, algorithmic technology factors, and social environment factors as key drivers of algorithm aversion. A comprehensive framework for analyzing the formation mechanism of algorithm aversion, based on the concept of "individual-algorithm-social environment," is extracted. Based on this framework, the study proposes research paths and coping strategies from three perspectives: theoretical research, technical research, and humanistic research. These recommendations aim to effectively address and mitigate algorithm aversion among social media users.

Key words: algorithm aversion, formation mechanism, personalized recommendation, coping strategy