Hengyi Fu

Introduction. This study examines algorithm aversion in digital library contexts, where users increasingly encounter algorithm-enabled services. Prior research in healthcare and finance shows that users often resist algorithmic systems even when they outperform human alternatives, yet little is known about how this phenomenon manifests in the digital libraries context. Method. As part of an ongoing project, we conducted semi-structured interviews with 18 users from two Chinese universities. Participants were selected through theoretical sampling. Interviews were analysed using constructivist grounded theory. Analysis. Open and axial coding of the first 18 interviews produced 42 initial concepts organised into 16 categories under five dimensions: algorithmic, platform, individual, task, and cognitive factors. Results. Preliminary findings suggest that algorithm aversion in digital libraries is multidimensional. Users reported concerns about inaccurate or irrelevant results, cumbersome interfaces, lack of familiarity, and limited trust in recommendations. These patterns suggest that aversion is shaped by both technical shortcomings and user-level factors, leading to frustration, avoidance, and reliance on manual methods. Conclusion(s). These early insights extend algorithm aversion research into a complex service environment, offering evidence that resistance arises from the interplay of technical, individual, and contextual conditions. The ongoing study will expand the dataset and employ DEMATEL to identify causal relationships among factors.

Purpose: Vaccine hesitancy has become a critical public health challenge, particularly during the COVID-19 pandemic, amplified by social media facilitating the spread of unverified claims. This study compares COVID-19 and non-COVID-19 vaccine questions posed on a leading USA health Q&A community, examining topical differences, user engagement and themes underlying vaccine hesitancy.Design/methodology/approach: Questions from Medical Sciences Stack Exchange were collected and pre-processed to identify vaccine-related queries. The analysis included 324 COVID-19 vaccine questions and 187 non-COVID-19 vaccine questions for topical differences. Vaccine hesitancy was examined in 75 COVID-19 and 44 non-COVID-19 vaccine questions. User engagement metrics were compared between the two categories.Findings: The study revealed distinct thematic concerns in COVID-19 versus non-COVID-19 vaccine questions, alongside substantial differences in user engagement metrics. Common strategies to address vaccine hesitancy may not be equally effective during public health crises. COVID-19 vaccine hesitancy primarily stemmed from concerns about rapid development, new biotechnologies and the benefits of post-infection vaccination. Non-COVID-19 vaccine hesitancy focused more on the safety of childhood immunization, manufacturing processes, allergy risks and vaccine-related disease myths. The high popularity but low answer acceptance rate of COVID-19 vaccine questions indicated a substantial unmet information need in addressing hesitancies.Originality/value: This study compares user concerns regarding established vaccines and those developed during a health crisis, focusing on factors contributing to vaccine hesitancy. It evaluates the extent to which these concerns are addressed on a health Q&A platform, offering insights into information needs and gaps in public understanding of vaccines during routine and emergency health scenarios.

Introduction. Despite the efficacy of vaccines, online misinformation contributes to vaccine hesitancy. This study investigates factors influencing the acceptance of answers to vaccine-related questions on social Q&A platforms, aiming to improve online vaccine communication. Method. We analysed 511 vaccine-related questions and 875 associated answers on the medical sciences stack exchange platform. Eleven variables were identified as potential predictors of answer acceptance, including citation practices, use of personal experiences, and recognition of asker's concerns. Analysis. Logistic regression was employed to determine which variables significantly predict the likelihood of an answer being designated as the ‘accepted answer’ by the asker. Results. Personal experience sharing, recognition of asker's concerns, citing data sources, and including calls-to-action emerged as significant predictors of answer acceptance. Notably, professional credentials had a modest effect, while factors such as the use of analogies or citing health authorities were not significant predictors. Conclusions. Our findings suggest that effective vaccine communication on social Q&A platforms should balance factual information with personal narratives, address user concerns empathetically, and provide clear, actionable advice. These insights can inform strategies for health communicators and platform designers to enhance the quality and acceptance of vaccine information online.

Purpose: This exploratory, descriptive study examines the factors that might influence the success or failure of online peer production communities during their inception stage. It focuses on community-building activities, the roles of users and the dynamics of user interaction, aiming to shed light on practices that could contribute to a community's success at the inception stage.Design/methodology/approach: By comparing two Stack Exchange Q&A communities with the same timelines but opposite outcomes during their beta testing phases, the research utilizes quantitative methods to categorize community activities, define user roles via k-means cluster analysis and examine interaction networks using social network analysis.Findings: Our findings suggest the successful Mathematics Q&A community exhibited several distinct practices during its inception, such as the utilization of both external and internal tools, the development of community-specific tutorials and the strategic use of flagging functions for moderation. Eight user roles were discerned, with roles like content editors, metadata curators and gatekeepers being particularly prominent in the successful community. Additionally, a more densely interconnected user network characterized by active participation was observed in the successful community.Originality/value: Concentrating on the inception stage of online communities, this study uncovers insights into the dynamics at play in the early life of peer production environments and provides empirical observations that may assist in shaping strategies for new online communities. It stands out by comparing communities within the same period to understand factors that may influence their early success.

Technology-based tools for mental health support, powered by certain algorithms, often become the first point of contact for people looking for mental health resources, especially in underserved communities. The different ways people understand and talk about their mental distress significantly affect the resources these tools recommend or exclude. Using the Google search symptom dataset, we looked into how people from socioeconomically different counties in Alabama search for mental health symptoms. We found that individuals in better-off areas mostly conducted searches using specific clinical terms, while those in less advantaged areas mostly used general mental health symptom terms. A similar trend was seen for physical symptoms, with people in disadvantaged areas often using general and pain-related terms, possibly indicating a lack of access to specialized care or wrong clinical treatments and raising concerns about opioid misuse. Also, counties with mainly African American populations had fewer mental health-related searches, suggesting there might be cultural or linguistic barriers. We discuss what these findings mean for designing algorithms focusing on health fairness. This study helps suggest how search engine algorithm designers can be aware of the societal factors affecting how people express distress and adjust search engine algorithms and interfaces to help reduce gaps in healthcare access.

Background: Considering social distancing requirements and isolation during the COVID‐19 pandemic and the widespread health mis/disinformation on social media, filling the pandemic‐related information gap is critical.Objectives: This study explored COVID‐19‐related question topics, topic correlation, evolving trends, and community interactions by leveraging data from a health question‐answering (Q&A) community (public and health professionals) over a 2‐year period.Methods: A total of 828 questions were collected and manually reviewed from a health Q&A and classified into 2492 codes. Question topic correlations, trends, and community interactions were generated using different statistical techniques.Results: Thirty‐seven topics were identified, and strong topic correlations were observed between the cognitive layer and other layers. These topics changed as the pandemic developed. Community interaction analyses revealed topics that were popular but received comparatively fewer responses.Discussion: The findings could illuminate the dynamic changes in people's interests and concerns related to COVID‐19 and their experiences of exchanging information in social Q&A communities.Conclusion: This study examined the practical significance of understanding topics of interest during a pandemic. The findings can inform the design of social health Q&A platforms that could reduce the risks of misinformation and disinformation.

Purpose: This study aims to investigate the nature and evolution of online communities in the early stages of their life cycles. The authors analyze the topics of discussions in an online community to identify issues related to community development. The authors also compare the topics of exemplary questions that founding members believed to be asked with the real questions based on members' information needs.Design/methodology/approach: The authors use Medical Sciences Stack Exchange, a health Q&A community of Stack Exchange, which requires four stages of development: definition, commitment, private beta and public beta. The authors collect postings of discussions and health questions in the first three stages, perform a content analysis of the postings and analyze the topics of discussions and health questions.Findings: The authors find that the topics of discussions evolved dynamically with the issues of community governance, role as a medical/health community, members and roles, content management, quality control and community design. The authors also find that the real questions included more specific and diverse issues than the exemplary questions that founding members expected.Originality/value: Theoretically, this study tests the community life cycle model in an online community that has explicit phase markers. The findings could shed light on community development and help prioritize issues to solve and decisions to make in its early stages. Additionally, this study focuses on the challenges and concerns in online health community building and solutions generated by collective efforts that could influence health communications in online communities.