User Comments Research Articles

A netnographic study of anabolic-androgenic steroid initiation videos on YouTube.

YouTube personalities (YP) have amassed significant followings by distributing health information. There is concern that YP involved in anabolic androgenic steroid (AAS) communities are sharing information on how people should approach initiating AAS use. This research sought to explore how YP are using their platform to distribute information on AAS initiation, what information is being presented and how people are responding. This mixed methods research utilised a netnographic approach, a method adapted from traditional ethnographic techniques for the study of online sociocultural phenomena. The research included a selection of YouTube videos (n = 37) and user comments replying to the videos (n = 584). YP addressed the positives and negatives of AAS initiation (76%; n = 28 and 76%; n = 28 respectively), provided harm reduction strategies (76%; n = 28), and advice on which AAS people should use (97%; n = 36). A qualitative thematic analysis of the comments found two overarching themes: 'Authenticity, capital and the influence of the YP on their audiences' and 'The evolving landscape of enhancement and health'. Viewers engaged with the videos due to the YPs lived experience using AAS, and YP influenced the health behaviours of viewers. Many believed that the videos were useful for health and lifestyle enhancement purposes. YP are disseminating diverse information on AAS initiation which is derived from their lived experience using AAS. Harm reduction is a key component of the YPs' message and viewers responded positively to a balanced and honest assessment of AAS initiation which is perceived to respect their personal autonomy.

Oral Diabetes Medication Videos on Douyin: Analysis of Information Quality and User Comment Attitudes.

Oral diabetes medications are important for glucose management in people with diabetes. Although there are many health-related videos on Douyin (the Chinese version of TikTok), the quality of information and the effects on user comment attitudes are unclear. The purpose of this study was to analyze the quality of information and user comment attitudes related to oral diabetes medication videos on Douyin. The key phrase "oral diabetes medications" was used to search Douyin on July 24, 2023, and the final samples included 138 videos. The basic information in the videos and the content of user comments were captured using Python. Each video was assigned a sentiment category based on the predominant positive, neutral, or negative attitude, as analyzed using the Weiciyun website. Two independent raters assessed the video content and information quality using the DISCERN (a tool for assessing health information quality) and PEMAT-A/V (Patient Education Materials Assessment Tool for Audiovisual Materials) instruments. Doctors were the main source of the videos (136/138, 98.6%). The overall information quality of the videos was acceptable (median 3, IQR 1). Videos on Douyin showed relatively high understandability (median 75%, IQR 16.6%) but poor actionability (median 66.7%, IQR 48%). Most content on oral diabetes medications on Douyin related to the mechanism of action (75/138, 54.3%), precautions (70/138, 50.7%), and advantages (68/138, 49.3%), with limited content on indications (19/138, 13.8%) and contraindications (14/138, 10.1%). It was found that 10.1% (14/138) of the videos contained misinformation, of which 50% (7/14) were about the method of administration. Regarding user comment attitudes, the majority of videos garnered positive comments (81/138, 58.7%), followed by neutral comments (46/138, 33.3%) and negative comments (11/138, 8%). Multinomial logistic regression revealed 2 factors influencing a positive attitude: user comment count (adjusted odds ratio [OR] 1.00, 95% CI 1.00-1.00; P=.02) and information quality of treatment choices (adjusted OR 1.49, 95% CI 1.09-2.04; P=.01). Despite most videos on Douyin being posted by doctors, with generally acceptable information quality and positive user comment attitudes, some content inaccuracies and poor actionability remain. Users show more positive attitudes toward videos with high-quality information about treatment choices. This study suggests that health care providers should ensure the accuracy and actionability of video content, enhance the information quality of treatment choices of oral diabetes medications to foster positive user attitudes, help users access accurate health information, and promote medication adherence.

Self-Diagnosis of Mental Disorders: A Qualitative Study of Attitudes on Reddit.

There is concern that a growing number of individuals, especially adolescents, are diagnosing themselves with mental disorders. However, there has been limited empirical research into this phenomenon: why it might happen, what the costs and benefits might be, and what the implications are for anyone who is experiencing distress. To address this, this study used reflexive thematic analysis to explore attitudes toward self-diagnosis of mental disorders as expressed on the discussion website Reddit. From 1195 user comments, five themes were generated: (1) There is tension over who is the expert in diagnosis; (2) Self-diagnosis is a route to self-understanding in an inaccessible system; (3) Teenagers on social media are the problem; (4) Self-diagnosis can become self-fulfilling, and (5) Now no one is believed. Together, these themes highlight that there is considerable anger, derision, and criticism targeted toward people who self-diagnose with mental disorders, and that this is particularly targeted toward adolescents who self-diagnose on or as a result of social media. The findings have important implications for understanding how to support and validate people, particularly adolescents, who (sometimes accurately) use diagnostic language to express how they are feeling.

Breakdown of the co-creation process on social media: developing the value co-washing framework and stakeholders dynamics

Purpose Social media platforms facilitate brand-consumer interactions by leveraging principles from nudging, value co-creation and social identity theories. This study aims to investigate how these interactions mask harmful practices and accelerate market access, perpetuating extreme consumerism. Specifically, the authors explore how value flow on social media, across distinct stakeholders, leads to value co-washing, revealing their collective unsustainable behaviours and related effects. Design/methodology/approach The authors use a mixed-methods approach, conducting content and sentiment analysis on nine TikTok videos featuring products from a leading Chinese company and analysing 19,816 user comments. Findings The value co-washing framework is developed across three building blocks: brands, creators and users. Findings uncover a paradigm shift in stakeholders’ dynamics, highlighting how social media collaborative engagements foster value co-washing. User involvement is categorized into three distinct clusters – brand lovers, saga creators and boycotters. The analysis identifies nine thematic patterns, including value co-creation, brand promotion, audience retention and calls for responsibility. Sentiment analysis reveals a dominance of neutral sentiments, reflecting a widespread unawareness and social adherence to value co-washing. Originality/value The proposed framework innovatively maps how distinct stakeholders contribute to extreme consumerism through value co-washing, providing foundational insights into the underlying mechanisms of consumer behaviour.

Research on Optimization Path of Higher Education Management Based on Collaborative Filtering Algorithm

In order to better realize the optimization of university education management, this paper puts forward the research on the optimization path of university education management under collaborative filtering algorithm. The optimization of higher education management is divided into several management directions. First, the teaching resource recommendation management, using collaborative filtering algorithm based on emotional tendency, analyzes the comments of users of teaching resources, and recommends the teaching resources to users next time according to the emotional tendency obtained from the analysis. Second, the library book recommendation management, constructing the characteristic model of library book borrowing users, uses collaborative filtering algorithm to generate nearest neighbors, calculating the similarity of users, and recommending books that meet their characteristics for users according to the similarity. Finally, the management of personalized course selection recommendation in colleges and universities, which uses the historical information of students to establish the evaluation matrix of course selection, searches the nearest neighbors according to the evaluation matrix, and produces the personalized course selection recommendation results of the students according to the nearest neighbors. According to the experiment, the recommended users can be satisfied with the recommended results in different directions of education management. After applying this method, the overall quality of college students has been significantly improved, the expenditure of education management in colleges and universities has been significantly reduced, and the optimization effect of education management in colleges and universities has been remarkable.

Public Perceptions of GAI: A YouTube Analysis of Sora

ABSTRACTText‐to‐video generative AI (GAI) tools are captivating the world with hyper‐realistic content. However, public perceptions of such technologies remain ambiguous. This paper investigates Sora, a new OpenAI offering, as an example to understand the public perceptions of GAI technology. The data comprises the 39 most popular YouTube videos about Sora and their most‐liked user comments. Through a SWOT analysis, we identify the multifaceted portrayals of Sora in the videos and the corresponding comments. Implications for future research are discussed.

Biased Social Media Debates About Terrorism? A Content Analysis of Journalistic Coverage of and Audience Reactions to Terrorist Attacks on YouTube

Social media are an important source of news during crises such as terrorist attacks. However, how news media and their audiences make sense of terrorism on social media is subject to bias, for example, given their differential treatment of terrorism by right-wing versus Islamist extremist perpetrators. In this study, we analyze how incident- and perpetrator-related characteristics of terrorist attacks are associated with bias in public debates about terrorism on YouTube. We focus on selectiveness in which attacks are covered (gatekeeping bias), how attacks are covered (presentation bias), and how audiences react to coverage (audience bias). Methodologically, we employ a manual and an automated content analysis of terrorism coverage by five international broadcasters on YouTube ( N = 643, 2018–2020) and related user comments ( N = 193,721). Our findings illustrate how sociocultural contexts shape news bias in terrorism coverage, both in the form of gatekeeping bias and presentation bias—but we conclude with less evidence for audience bias in public reactions to terrorism, at least on social media. Consequently, journalists should critically question working routines in covering crises to avoid reinforcing power imbalances, especially those from Western contexts.

Text Analytics on YouTube Comments for Food Products

YouTube is a popular social media platform in the contemporary digital landscape. The primary focus of this study is to explore the underlying sentiment in user comments about food-related videos on YouTube, specifically within two pivotal food categories: plant-based and hedonic product. We labeled comments using sentiment lexicons such as TextBlob, VADER, and Google’s Sentiment Analysis (GSA) engine. Comment sentiment was classified using advanced Machine-Learning (ML) algorithms, namely Support Vector Machines (SVM), Multinomial Naive Bayes, Random Forest, Logistic Regression, and XGBoost. The evaluation of these models encompassed key macro average metrics, including accuracy, precision, recall, and F1 score. The results from GSA showed a high accuracy level, with SVM achieving 93% accuracy in the plant-based dataset and 96% in the hedonic dataset. In addition to sentiment analysis, we delved into user interactions within the two datasets, measuring crucial metrics, such as views, likes, comments, and engagement rate. The findings illuminate significantly higher levels of views, likes, and comments in the hedonic food dataset, but the plant-based dataset maintains a superior overall engagement rate.

Lessons learned from Dread darknet communities: How and why are fraudsters targeting the elderly to be victims or accomplices?

AbstractResearch SummaryWe examined darknet user discussions on the Dread forum to identify key themes and emerging topics in fraud planning, focusing additionally on elderly victimization. Using the conceptual framework of criminogenic learning to study the process of fraud planning in darknet communities of practice, we analyzed the content of original user posts (n = 818) and comments (n = 1365) collected from the Dread forum. We identified nine unique fraud categories, of which payments‐related fraud was the most‐discussed fraud category and accounted for 25% of original user posts. We further found our sampled forum content to be dominated by the theme of learning or knowledge sharing in eight of the nine fraud categories, which appeared in more than 44% of original user posts. Our content analysis revealed specific insights into why and how darknet forum users may target people, including the elderly subpopulation, for different types of fraud. Overall, our research demonstrates the diversity of opinion and knowledge sharing among darknet forum users in relation to planning and committing fraud against the elderly, views on who is a suitable target and why, and what veteran and aspiring fraudsters consider to be important information for success in fraud.Policy ImplicationsUtilized independently or as part of a multistage strategy of darknet data analysis, our research method can be used to monitor criminally active darknet forums for current and emerging themes in fraud against general populations and vulnerable subpopulations, such as the elderly, and to develop strategies to identify, disrupt, or destroy hubs of criminal planning and knowledge sharing. Our study additionally informs policy makers of certain elder‐specific vulnerabilities that might be addressed by more up‐to‐date elder cybercrime awareness campaigns and initiatives. Our findings also highlight the presence of insider threats that may inform discussions on how health and financial institutions can better regulate these cybersecurity risks.

Harnessing Hope in Digital Spaces for Health Equity: How Group Comparison Information and Supportive Comments Influence Bone Marrow Donor Intentions for African Americans.

Pursuing health equity necessitates recognizing health disparities that disproportionately impact disadvantaged groups and eliminating their barriers to essential health resources. Interactive digital technologies-specifically, popular social media platforms such as blogs and social networks-can be leveraged to engage underserved minority populations in collective social action aimed at addressing key determinants of health disparities and promoting equitable health outcomes. The present research focuses on the plight of African Americans-a minority group facing significant health disparities. Particularly in the domain of bone marrow donation, African Americans remain the group least likely to find a matching donor. Guided by the social comparison framing literature, we conducted an online experiment to investigate how group comparison information (GCI) emphasizing group-based disparities and supportive user comments on social media platforms influence African Americans' intentions to join a bone marrow registry. In doing so, we considered hope as a mediator and group identification as a moderator. Results based on a conditional process analysis showed that GCI led to greater bone marrow donor intentions in the presence of supportive comments through elicitation of hope, particularly among those low in group identification. The current findings demonstrate that it is important to consider the role of supportive message environments and group identification when addressing health disparities with GCI. Theoretical and practical implications are discussed.

Research on the product language construction of China’s twisted porcelain cultural heritage

This paper studies the product language construction of the twisted porcelain cultural heritage. Through field research, we collected and sorted out samples of twisted porcelain products, explored the product language characteristics of twisted porcelain from multiple aspects such as production process, product shape, and product color, interpreted cultural value, captured potential connotations, extracted representative words from user comments, quantified the relationship between users and twisted porcelain culture, realized the construction and transmission of traditional cultural language information, conveyed the traditional cultural information of the product to users, and promoted the sustainable dissemination and development of this cultural heritage. The research results show that after mining and extraction at the level of twisted porcelain characteristics, the core language constructs the cultural expression of twisted porcelain products, which is more in line with the needs of the market and users, and has the potential to be developed and disseminated using the language generation of cultural heritage products.

The Impact of Journalistic Cultures on Social Media Discourse: US Primary Debates in Cross-Lingual Online Spaces

This cross-lingual project examines how social media posts of Spanish- and English-language media impact incivility in user comments during the 2020 primary political debates in the United States. We analyzed Facebook posts of news organizations that hosted the debates and used a state-of-the-art machine-learning model to analyze the corresponding comments. Our findings reveal distinct journalistic cultures on the post-level: English-language media are significantly more likely to use interpretation while Spanish-language media employ more audience-engagement and factual reporting strategies. We argue that in order to understand incivility in social media discourse during political debates, we need to consider journalistic cultures: While interpretative reporting explains lower levels of incivility in the English-language discourse, factual reporting explains lower levels of incivility in the Spanish-language discourse. We suggest that we need to consider how features of news reporting (textual and visual) impact discourse quality directly but also indirectly via emotional arousal in comments.

Revisiting Sentiment Analysis for Software Engineering in the Era of Large Language Models

Software development involves collaborative interactions where stakeholders express opinions across various platforms. Recognizing the sentiments conveyed in these interactions is crucial for the effective development and ongoing maintenance of software systems. For software products, analyzing the sentiment of user feedback, e.g., reviews, comments, and forum posts can provide valuable insights into user satisfaction and areas for improvement. This can guide the development of future updates and features. However, accurately identifying sentiments in software engineering datasets remains challenging. This study investigates bigger large language models (bLLMs) in addressing the labeled data shortage that hampers fine-tuned smaller large language models (sLLMs) in software engineering tasks. We conduct a comprehensive empirical study using five established datasets to assess three open-source bLLMs in zero-shot and few-shot scenarios. Additionally, we compare them with fine-tuned sLLMs, using sLLMs to learn contextual embeddings of text from software platforms. Our experimental findings demonstrate that bLLMs exhibit state-of-the-art performance on datasets marked by limited training data and imbalanced distributions. bLLMs can also achieve excellent performance under a zero-shot setting. However, when ample training data is available or the dataset exhibits a more balanced distribution, fine-tuned sLLMs can still achieve superior results.

The Ethical Implications of Prosocial Synthetic Resuscitation: Analysing User Comments to a Deepfake Campaign Addressing Intimate Partner Violence

In recent years, the ‘digital resurrection’ of the deceased through Deepfake technology has become a reality accessible to the general public and has raised profound ethical discussions in society. While the academic literature focuses primarily on regulatory and political implications, the public reception of digital resurrection remains largely unexplored. Furthermore, much of the public discourse around Deepfakes centres on their malicious applications, despite the potential for digital resurrection to serve prosocial purposes. This study focuses on an Israeli campaign called ‘Listen to My Voice’ (2021), which ‘resurrected’ women murdered by their partners, having them ‘tell’ the stories of the abusive relationships that preceded their deaths. An analysis of viewer responses shows that the campaign elicited strong reactions straddling the line between technology and ethics. Quantitatively, most comments supported the use of Deepfakes, given the importance of the cause and the impactful messaging. Qualitatively, two key themes emerged: one technical, indicating that visceral realism sparked both emotional connections and fundamental misunderstandings of the technology itself. The second highlighted an ethical tension between desires to protect autonomy and justifications for posthumously using an individual’s likeness for positive aims. The findings suggest Deepfakes’ technical attributes, particularly tangibility, significantly shape its ethical reception—a phenomenon we term ‘the ethics of technological misunderstanding’.

The Role of Networked Narratives in Amplifying or Mitigating Intergroup Prejudice: A YouTube Case Study

This purpose of this research is to understand the role of networked narratives in social media in modulating viewer prejudice toward ethnic neighborhoods. We designed experimental videos on YouTube based on intergroup contact theory and narrative frameworks aimed at (1) gaining knowledge, (2) reducing anxiety, and (3) fostering empathy. Despite consistent storytelling across the videos, we observed significant variations in viewer emotions, especially in replies to comments. We hypothesized that these discrepancies could be explained by the influence of the surrounding digital network on the narrative’s reception. Two-stage research was conducted to understand this phenomenon. First, automated emotion analysis on user comments was conducted to identify the varying emotions. Then, we explored contextual factors surrounding each video on YouTube, focusing on algorithmic curation inferred from traffic sources, region, and search keywords. Findings revealed that negative algorithmic curation and user interactivity result in overall negative viewer emotion, largely driven by video placement and recommendations. However, videos with higher traffic originating from viewers who had watched the storyteller’s other videos result in more positive sentiments and longer visits. This suggests that consistent exposure within the channel can foster more positive acceptance of cultural outgroups by building trust and reducing anxiety. There is the need, then, for storytellers to curate discussions to mitigate prejudice in digital contexts.

Analisis Sentimen Persepsi Masyarakat Terhadap Timnas Indonesia U-23 dalam AFC-23 Asian Cup 2024 Pada Media Sosial X Menggunakan Metode Naïve Bayes Classifier

Social media Twitter, or now called X, not only serves as an effective communicative medium, but it can also make it easier for people to find information or interact directly with other X users, and become a container for holding opinions. X users in Indonesia often express their opinions on all areas including sports, especially the Indonesian Football Team, which has recently been heatedly discussed because of the talent of all the players who succeeded in attracting the attention not only of football fans, but of the entire Indonesia public at the AFC ASIA CUP U 23 2024 competition held in the country of Qatar. The study uses the Naive Baiyes classification method to analyze the sentiment of comments and comments on Social Media X against each match. Sentiment data from user comments on social media X is collected and classified as positive, and negative. The final results of sentimental analysis using the naïve bayes method showed accuracy values of 73.94% out of 972 datasets, with 143 predicted negative data and 828 predictable positive data. So it can be concluded that most of the public's perception is positive because they see the struggle and performance of the Indonesian National Football Team in making the nation proud.

Yesterday once more: collective storytelling and public engagement with digital cultural products on the music streaming platform

Drawing on narrative transportation theory, we propose that when people consume a cultural product, they consume their emotions and memories through collective storytelling. Such emotions and memories are amplified by user comments on social media, enhancing the product’s influence and triggering audience engagement. We collected public data from NetEase Cloud Music—a major music streaming platform in mainland China—to investigate how the emotions and memories expressed in user comments influence the songs’ impact. Findings indicate that autobiographical narratives and negative emotions in user comments significantly boost a song’s influence. Moreover, user comments are particularly effective in promoting emerging artists with limited resources compared to their more established counterparts. This study advances the narrative transportation theory by emphasizing the role of active audiences and collective storytelling. Empirically, it extends the existing literature on the factors influencing cultural products on music streaming platforms in a non-Western context.

Using Psychological Reactance and Bandwagon Cues to Explain High School Sports Coaches’ Resistance to Concussion Communication Policies

Sport-related concussions (SRC) are a significant health risk for athletes amalgamated by issues with underreporting. Coaches play a pivotal role in promoting safety and health, but coach-player SRC communication efforts have been insufficient at improving players’ reporting behaviors. This study considers coaches’ responses to a proposed mandate to engage in SRC communication with athletes via psychological reactance theory (PRT), with bandwagoning cues positioned as a form of social proof that heightens or mitigates persuasion. Two hundred five coaches from 20 distinct sports were exposed to one of four messages addressing a proposed SRC management policy mandating biweekly coach-player communication about SRCs. The message variations incorporated different bandwagon cues (positive, negative, absent, mixed), manipulated through user comments and emoji reactions on a simulated social media environment. Results demonstrate direct and mediated relationships between exposure to negative bandwagon cues, bandwagon perceptions, freedom threat, reactance, and opposition intentions. Our findings provide valuable insights into the dynamics of coaches’ attitudes towards SRC communication in a social media context, bridging gaps in the literature and introducing PRT to the SRC domain.

Going Viral: Sharing of Misinformation by Social Media Influencers

This paper tackles the issue of health and well-being misinformation, the dissemination of false or misleading information related to medical treatments, diseases, mental health, nutrition, exercise or general well-being, propagated by social media influencers. It investigates the virality of misinformation posts by TikTok and Instagram influencers exploring users’ appraisals and their sharing tendencies. Grounded in social influence and cognitive appraisal theories (CAT), three online experimental studies dissect the dynamics of virality, user comments and their effects on perceived deception, parasocial interaction and sharing intent. The results of Study 1 demonstrate heightened post virality reduces perceived deception, fostering stronger parasocial connections and sharing intentions. Conversely, lower virality levels heighten deception perceptions. In Study 2, critical comments are shown to enhance deception in high virality posts. Whereas supportive comments are shown to enhance the indirect effect of low virality posts on sharing intentions, via deception and parasocial interaction. The study contributes by demonstrating how social influence theory and CAT together explain how social media influencer misinformation posts based on their virality and user responses are likely to be shared and what consumer appraisals explain this effect. It provides directions of how marketers can tackle this issue.

Classifying Application User Comments Using the Improved K-Nearest Neighbor and BM25F Weighting Methods

Classifying Application User Comments Using the Improved K-Nearest Neighbor and BM25F Weighting Methods