Facebook Activity Research Articles

From entry to transformation: exploring AI integration in teachers’ K–12 assessment practices

ABSTRACT This study explores the role of artificial intelligence (AI) in K–12 student assessment practices, focusing on educators’ use of AI tools. Through content analysis of active Facebook groups dedicated to AI in education, the authors examined how educators integrate AI into assessment across various grade levels and subjects. Using the Technology Integration Model, they analysed posts to evaluate the use of AI across subjects and grade levels. The findings revealed a progression from initial interest and experimentation with AI to more advanced, adaptive uses in assessment creation and grading. Educators reported positive experiences with AI-driven tools, especially in automating tasks and gamifying assessments. However, challenges such as ensuring reliable, context-specific feedback persist. Overall, the research highlights AI’s potential to enhance student-centred assessment practices and provides insights for educators, administrators and policymakers on AI’s evolving role in education.

The Relative Probability of Facebook Friendship in the United States

We mapped Facebook’s Social Connectedness Index (SCI) between adjacent counties in the Contiguous 48 U.S. States. The index is calculated as the number of Facebook friends between counties, divided by the product of active Facebook users in the two counties. The results follow regional science principles that tell us that fewer flows may occur across political (administrative) borders such as state boundaries, and between economic zones, including transition zones between metropolitan areas and hinterland boundaries. We also found low connectivity between adjacent counties that are divided by interstate highways and low connectivity within densely populated areas. High connectivity is found in rural areas, and areas of cultural significance, such as highly African American regions in the U.S. South and isolated regions in Appalachia.

Strained ties in plurinational states: Analysing the social network divide between Canada's two solitudes

AbstractPatterns of social networks in plurinational societies remain understudied in the inter‐group contact literature. In Canada, a significant number of authors have studied political divides between Quebecers and other Canadians, but little is still known about the scope and determinants of the social disconnect between both communities. This paper provides the first quantitative measure of the social networking gap between Canada's ‘two solitudes’ using behavioural data. Our analysis mobilises the Facebook Connectedness Index (2020), which provides district‐level connectivity estimates between all active Facebook users in Canada. Results obtained via linear regression models show that population size, distance and self‐isolation patterns cannot account for the significant connectivity gap between Quebec and the rest of Canada. However, bilingualism in Quebec is positively correlated with higher inter‐regional networks, suggesting asymmetrical linguistic exchanges. These results empirically confirm decades of anecdotal accounts about the existence of a social disconnect between Canada's French‐ and English‐speaking communities.

Multidimensional Cross-Country Measurement of Climate Change Perceptions

ABSTRACT Understanding public perceptions of climate change is crucial for more targeted communication and better-informed policymaking. Moreover, the perceptions of climate change encompass different aspects and thus need to be defined and measured using a multidimensional approach. In this paper, we introduce the Climate Perceptions Index (CPI), a composite measure that comprehensively assesses the public perceptions of climate change, based on a sample of over 100 thousand active Facebook users across 107 countries. We construct the CPI as an aggregate of three distinct dimensions that quantify the awareness of climate change, risk perception, and commitment to action. The results show that the extent of climate change perceptions varies significantly across countries and across dimensions. Countries with the highest and lowest scores globally can be found in almost all regions, and across all levels of socio-economic development. Furthermore, we analyze the relationships between the CPI dimensions and find that the influence of risk perception on commitment to action is the strongest at the lowest levels of climate change awareness. This highlights the possibility of climate change risk normalization at higher levels of awareness, and further shows that effective policies and strategies must not only focus on raising knowledge about climate change, but also overcome the normalization of its threats.

Implementing eHealth-based behaviour change support within a nutrition intervention trial improves adherence to study-related behaviours in healthy young adults

Nutrition intervention trials play a key role in informing clinical and dietary guidelines. Within these trials, we need participants to change their behaviours; however, researchers seldom systematically consider how to support participants with these changes, contributing to poor adherence. Here we evaluate how using a behaviour change framework to develop support within a dietary intervention impacts young adults’ adherence to required trial behaviours. In the Protein Diet Satisfaction (PREDITION) trial, 80 young adults were randomised to a flexitarian or vegetarian diet for 10-weeks to investigate the psychological and cardiometabolic effects of moderate lean red meat consumption as part of a balanced diet(1). To understand these outcomes, it was key that participants within the trial (i) ate a healthy, basal vegetarian diet (excluding meat, poultry, and fish not provided by research team) and (ii) reported their dietary intake daily on a smartphone application (required to evaluate intervention compliance). To enhance adherence to these behaviours the Nine Principles framework was used to develop behaviour change support (BCS)(2). Key components of the BCS included access to a dietitian-led Facebook group, text reminders, and food delivery. Effectiveness was measured using the following analyses of the 78 participants who completed the study: pre-post change in targeted dietary habits over time using a subscore of the Healthy Diet Habits Index, adherence score to reporting over 10-weeks, Facebook group engagement, and impact evaluation. Analysis included linear imputation modelling, t-tests, and chi-square analysis. The total Healthy Diet Habits Index subscore out of 16 significantly increased from baseline to week 10 (10.6 ± 2.6 to 11.2 ± 2.6, p = 0.011), demonstrating maintenance of a healthy diet. Overall adherence to reporting was high across the 10 weeks, with the total population mean reporting score 90.4 ± 14.6 out of a possible 100. This strengthens study validity, allowing us to confidently report if participants complied with study requirements of consuming the intervention protein (red meat or plant-based meat alternatives) on top of a basal vegetarian diet. Although relatively low active Facebook engagement was observed (on average <1 ‘react’ per post), most participants agreed the text messages and Facebook groups supported them to adhere to recording (63%) and eating healthily (60%), respectively. This is the first study to provide an example of how a framework can be used to systematically develop, implement, and assess BCS within a nutrition trial. This appears to be a promising way to enhance adherence to study-related behaviours, including the burdensome task of reporting dietary intake. We believe this has great potential to improve research validity and decrease resource waste, not only for the PREDITION trial but in future dietary intervention trials.

Longitudinal Patterns of Online Activity and Social Feedback Are Associated with Current and Perceived Changes in Quality of Life in Adult Facebook Users

The present study explored how sharing verbal status updates on Facebook and receiving Likes, as a form of positive social feedback, correlate with current and perceived changes in Quality of Life (QoL). Utilizing the Facebook Graph API, we collected a longitudinal dataset comprising status updates and Likes received by 1577 adult Facebook users over a 12-month period. Two monthly indicators were calculated: the percentage of verbal status updates and the average number of Likes per post. Participants were administered a survey to assess current and perceived changes in QoL. Confirmatory Factor Analysis (CFA) and the Auto-Regressive Latent Trajectory Model with Structured Residuals (ALT-SRs) were used to model longitudinal patterns emerging from the objective recordings of Facebook activity and explore their correlation with QoL measures. Findings indicated a positive correlation between the percentage of verbal status updated on Facebook and current QoL. Online positive social feedback, measured through received Likes, was associated with both current QoL and perceived improvements in QoL. Of note, perceived improvements in QoL correlated with an increase in received Likes over time. Results highlight the relevance of collecting and modeling longitudinal Facebook data for the investigation of the association between activity on social media and individual well-being.

Abstract P366: Targeted Facebook Ads Increased Engagement Among Rural Communities During Pre-Recruitment Phases of the Mississippi Core of the Risk Underlying Rural Areas Longitudinal Study

Background: Data is sparse regarding the utility of social media in engaging southern, rural communities to participate in health research. Community Advisory Board (CAB) members of the Mississippi (MS) Core of the Risk Underlying Rural Areas Longitudinal (RURAL) Study suggested the use of the Facebook social media platform to promote the RURAL Study. During the pre-recruitment phase of engagement (6 months prior to initiation of examinations and 3 months prior to official recruitment), RURAL investigators used Facebook posts to help increase awareness of and interest in the RURAL Study. Objective: To use Facebook metrics to identify the types of posts that garnered the most engagement. Methods: The Principal Investigator of the RURAL Study’s MS Core collaborated with the University of MS Medical Center Social Media Coordinator to implement a Facebook page for the Study. The MS team implemented social media posts to promote the RURAL Study. Data on the number of post reactions, comments, and shares, as well as the reach of organic and paid ads, were collected to identify posts with highest engagement. Findings: In Oktibbeha County, the greatest number of Facebook profile visits occurred October 15-November 7, 2022. Facebook posts had the greatest reach between June 24-27 and September 19-30, 2022. Paid ads (boosted posts) produced greater reach than organic ads. In addition, post reactions positively correlated with the number of people reached. Posts with the greatest number of reactions were related to RURAL-sponsored events (e.g., the County Health Fair and the RURAL Open House event). There was significant Facebook activity around popular community events during which the RURAL Study was promoted (e.g., Fannie Lou Hamer event, the Open House event, and the Black Jack Morning Church Service). Although MS Core of RURAL investigators participated in interviews on local TV media prior to recruitment, organic Facebook ads that included links to videos of these interviews did not receive many reactions on the MS Core of the RURAL Study’s Facebook page. Conclusion: Using social media may be an effective tool for reaching communities during the pre-recruitment phase of engagement. Our findings support that paid ads that targeted specific audiences were more effective than organic ads. Facebook posts that documented RURAL’s participation and engagement in community events were more effective than posts about RURAL’s promotion in broadcast media.

Online brand community engagement and brand evangelism: the role of age, gender and membership number

PurposeThis study aims to investigate the critical role of online brand community (OBC) engagement and brand evangelism: the role of age, gender and membership number.Design/methodology/approachTo achieve the study’s objective, 303 active Facebook OBC users were surveyed. The AMOS structural equation modelling (SEM) is used to test the hypothesised model.FindingsThe results indicate that OBC engagement improves brand evangelism. Specifically, the results show that age, gender, and OBC membership number moderate the relationship between OBC engagement and brand evangelism. Interestingly, this study found that female, younger and low-OBC follower consumers more significantly contribute to nourishing brand evangelism than male, older and high-OBC follower counterparts.Originality/valueOBC engagement is validated as a key brand evangelism driver, further substantiating its role as a crucial strategic metric. Moreover, age, gender and OBC membership number as moderating factors in the association between OBC engagement and brand evangelism (word of mouth referral, brand defence and future purchases) have been verified. Although the findings suggest that improved OBC engagement contributes to evangelism, this effect transpires more significantly among female, younger and low-OBC followers than male, older and high-OBC followers.

Critical Discourse Analysis on Parental Language Ideologies of Bilingual and Multilingual Child-Rearing and Language Education Using Facebook and Internet Forums

This study examines the computer-mediated discussion topics of parents who raise bilingual and multilingual children in four active Facebook and Internet forums, and investigates how the language ideologies embedded in the multiple languages being used in these forums are expressed. In this study, 179 data points, including users’ posts and thread comments, were collected to identify the most frequently discussed topics as part of my description of the database, in order to identify parental ideologies by using values analysis. The five most-discussed topics were selected to make a critical discourse analysis on the narratives to understand the language ideologies regarding the use of multiple languages, and regarding what users of the groups are saying specifically about the languages when analyzing metalinguistic discourses. This study found the most recurrent language ideologies that parents expressed on these online forums were supporting bilingualism/multilingualism, and claim that bilingualism/multilingualism is advantageous. Parents also demonstrate language ideologies supporting keeping languages separate, such as following the one parent one language (OPOL) method, using the minority language at home, and so on. A detailed values analysis with illustrative sample messages from the online posts and comments also more specifically shows the recurrent language ideologies identified, and parents’ views underlying their narratives on their posts and thread comments.

Facebook likes and corporate revenue: testing the consistency between attitude and behavior

Whether a person’s attitude is predictive or consistent with their behavior is a topic that has generated much research in the literature. The current study attempts to address this research question using a big data approach in a social media advertising context. Both attitude (i.e. clicking the “like” button for a company’s Facebook posts) and behavior (i.e. purchasing products from the company) are measured in a naturalistic setting. The goal is to examine whether Facebook “likes” on companies’ posts are significantly related to those companies’ revenue. The authors estimate panel models by using nearly eight years of data containing the S&P 500 companies’ Facebook activities in conjunction with their financial performance information. The results suggest that the number of Facebook “likes” is positively associated with revenue across the models. The study concludes with specific theoretical and practical implications and limitations.

Learning to bid and rank together in recommendation systems

AbstractMany Internet applications adopt real-time bidding mechanisms to ensure different services (types of content) are shown to the users through fair competitions. The service offering the highest bid price gets the content slot to present a list of items in its candidate pool. Through user interactions with the recommended items, the service obtains the desired engagement activities. We propose a contextual-bandit framework to jointly optimize the price to bid for the slot and the order to rank its candidates for a given service in this type of recommendation systems. Our method can take as input any feature that describes the user and the candidates, including the outputs of other machine learning models. We train reinforcement learning policies using deep neural networks, and compute top-K Gaussian propensity scores to exclude the variance in the gradients caused by randomness unrelated to the reward. This setup further facilitates us to automatically find accurate reward functions that trade off between budget spending and user engagements. In online A/B experiments on two major services of Facebook Home Feed, Groups You Should Join and Friend Requests, our method statistically significantly boosted the number of groups joined by 14.7%, the number of friend requests accepted by 7.0%, and the number of daily active Facebook users by about 1 million, against strong hand-tuned baselines that have been iterated in production over years.

Resisting right-wing populism in power: a comparative analysis of the Facebook activities of social movements in Italy and the UK

ABSTRACT This paper aims to present a comparative study of the civil society reaction to right-wing populism in power through social media, by looking at cases in Italy and the United Kingdom. The research question is how social movements are implementing a process of reactive political identity construction – i.e. political identification – and a political counter-strategy by opposing right-wing populism in power through their Facebook official accounts. It implements a mixed-method research design with in-depth semi-structured interviews and a two-step quantitative text analysis based on Topic Model and Dictionary Method.

Users' Motivations for Facebook Unfriending During the COVID-19 Pandemic: Survey Study.

Social networking sites (SNSs) such as Facebook have been central to the global exchange of health-related information throughout the COVID-19 pandemic, but during this time, increased web-based interactions proved to be a source of stress and conflict for many SNS users. Prior research suggests that many users have engaged in significant boundary regulation during this period, using behaviors such as unfriending to refine and reorient their social networks in response to pandemic-related information. This study aimed to examine Facebook unfriending during and in relation to the first year of the pandemic to better understand how SNS users have managed and maintained their social networks around the COVID-19 pandemic. On the one hand, unfriending may be motivated by an attempt to protect the utility and accuracy of a user's informational environment. On the other hand, it may be motivated by a desire to tune out alternative viewpoints and opinions. Both motivations may have significant implications for public health discourse and outcomes. A sample of 824 active Facebook users (drawn from a representative survey of 1000 American adults) was analyzed using a series of logit regression models. Survey respondents were selected using a stratified quota sampling approach to ensure a representative sample of the US population. Balanced quotas were determined (by the region of the country) for sex, age, race, ethnicity, and political affiliation. In total, 31.7% (261/824) of active Facebook users unfriended at least one account over COVID-19 pandemic-related posts during the first year of the pandemic. The most common reasons for unfriending included "making political comments about COVID-19" (191/824, 23.2%) and "posting information that was inconsistent with public health guidelines" (162/824, 19.7%). As hypothesized, reliance on Facebook for COVID-19 pandemic-related news and information was associated with a greater likelihood of unfriending, particularly in response to information that was inconsistent with public health guidelines. Political factors (particularly partisan intensity) were also predictive of unfriending, especially in the case of COVID-19 pandemic-related disagreements. Both information utility concerns and political factors were associated with a greater likelihood of COVID-19 pandemic-related unfriending, although the magnitude of the effects associated with utility appears to be greater. Although utility-motivated unfriending may lead to more reliable health information experiences for some SNS users, the tendency of consumers to assess accuracy and credibility on the basis of partisan predilections obscures this finding and warrants further consideration.

Impacts of Facebook on Reading Habits of Higher Secondary Students: Stakeholders’ Perspectives

Facebook is the most popular network site among young Bhutanese people. Students spend a substantial amount of time checking Facebook and chatting with friends, which could be detrimental to their reading habits. The aim of this study was to investigate the impacts of Facebook on the reading habits of a higher secondary school students in Bhutan. A mixed method, specifically a convergent design, was adopted for the study, and it was guided by the Pragmatism worldview. Quantitative data were gathered from 274 participants and qualitative data were collected from 10 interviewees and one focus group discussion. Participants included teachers, parents, and students from a higher secondary school in Thimphu Thromde. A descriptive and inferential statistical following thematic analysis was used for quantitative data and the qualitative data were analyzed along the same themes. The findings of the study revealed that use of Facebook positively affected students’ reading habits. Students developed positive attitudes towards reading, after they were exposed to inspiring and interesting articles. However, the study also highlighted the negative impacts of Facebook on students’ reading habits. Further, the study revealed that students have become complacent in reading owing to Facebook activities. The findings of the study may help different stakeholders understand the importance of creating awareness about media literacy to enable students to choose authentic reading materials on Facebook. This study recommends the Ministry of Education to develop policy and provide media literacy education to students to ensure they consume media information judiciously.

Socialization, Education, and Learning for the Internet (SELFI): A Pilot RCT of a Social Media Skills Group Program for Autistic Adults.

Many autistic adults report preference for computer-mediated communication and social media use. Despite many benefits to online socialization, there are many challenges including anxiety and cyber-victimization. To date, support is limited related to helping autistic adults with safe and effective internet use. The purpose of this study was to evaluate the feasibility, acceptability, and preliminary efficacy of the novel SELFI program. This pilot study utilized a randomized controlled trial design. A total of 25 autistic adults enrolled in the study and were randomized to the nine-week SELFI program or a waitlist control condition. Feasibility assessed enrollment, attrition, and fidelity of delivery. Acceptability examined attendance and feedback from participants and peer mentors. Efficacyevaluated change in Facebook activity, social media utility/anxiety, and individualized goals. Regarding feasibility, the recruitment goal was met within one month, there was limited attrition, and therapists delivered the program with high fidelity. Participants attended a majority of scheduled sessions and feedback from participants reflected high levels of agreement with several facets of the program. Compared to the control group, more participants assigned to the SELFI condition were perceived by autistic and non-autistic raters as having improved Facebook activity. SELFI participants also reported reduced difficulty meeting their individualized goal. Findings support the piloted SELFI program as feasible and acceptable with signals of preliminary efficacy. This study establishes an exciting foundation regarding an innovative social media skills program, however more research is necessary.

Socialization, education and learning for the internet (SELFI): development and pilot of a social media skills program for autistic adults

PurposeThe purpose of this paper is to describe the development and pilot of a novel program to support autistic adults with social media use. Social media use among autistic adults has been associated with increased happiness and closer friendships. However, autistic adults are at risk of social media challenges such as cyber-victimization. To date, no programs exist that specifically support autistic adults with safe and effective social media use. The primary aim of this study was to develop and pilot test a novel social media skills program for feasibility and acceptability. A secondary aim was to explore changes in Facebook activity as a proxy for online social engagement.Design/methodology/approachThe socialization, education and learning for the internet (SELFI) program was developed by identifying relevant target skills, goals, considerations and evidence-based strategies to help autistic adults with online social communication. The developed SELFI program was then pilot tested with six autistic male young adults to explore initial feasibility (fidelity of delivery, attendance) and acceptability (attrition, enjoyment and programmatic feedback). Facebook data were collected to explore differences in the frequency of likes/reactions, comments and posts after completion of the program.FindingsEach participant completed all procedures and attended every session. There was no attrition from the program. The average fidelity score for recorded sessions was 85.1%. Most participants endorsed agreement with program enjoyment and that it helped them maintain current relationships. Participant feedback provided valuable information regarding areas of strength and areas that needed improvement. Each participant who completed the study demonstrated increased Facebook engagement across likes/reactions, comments and posts.Research limitations/implicationsLimitations include a small all-male sample, exclusion of adults with intellectual disabilities and adolescents and dependence on Facebook data. Additionally, the developed program did not incorporate feedback from relevant stakeholders, including autistic adults.Originality/valueTo the best of the authors’ knowledge, this is one of the first studies to explore support for autistic adults specifically related to social media use.

Defending “Canadian Energy”: Connective Leadership and Extractive Populism on Canadian Facebook

ABSTRACT This paper provides a content and discourse analysis of the Facebook activity of pro-fossil fuel advocates in Canada, including industry-funded and independent groups. It explores how the fossil fuel industry and its allies have reacted to growing public demand for climate action by adopting a “subsidised public strategy,” in which elites provide stakeholders with resources to participate in an organized campaign. We argue that pro-oil advocacy groups utilize their Facebook pages to exercise “connective leadership,” in which key social media accounts strategically shape internet-enabled social movements by promoting mobilization opportunities, connecting supporters with diverse content, and reframing that content to communicate a coherent “extractive populist” storyline positioning “Canadian Energy” as a national public good under attack by foreigners and elites. We thereby provide insight into how industry and its allies leverage platform affordances to promote populist politics on Facebook to defend industry in the face of the climate crisis.

A social mapping of Swedish environment-focussed Facebook groups: The principles, methods and implementation of a mapping project

Aiming to contribute to the sparse literature on social mapping methodologies, this article elaborates on the principles, procedures and research design of a social mapping project. It first demonstrates the design and implementation of a mapping project of active Facebook groups in Sweden that focus on the environment, by elaborating on the project’s mapping principles and procedures. The developed mapping methodology integrates a variety of data gathering and analysis methods (survey research, online search, big data analysis, content analysis), applying the principle of triangulation, to ensure the completeness of the mapping. It also combines small-scale and big data analysis, addressing the advantages and limitations of these methods, in the framework of the specific research. In the second part, the article gives the overview of the social map of the 152 identified environment-focussed Facebook groups, connected to Sweden, during 2019–2020, presenting their basic features, networks and environmental focal points. In their majority these groups adopt ecocentric positions towards the environment and openly critique anthropocentrism, often being connected to civil society and activist networks, concerning environmental issues and action. These findings point to the potential of coordinated environmental action, being communicated and mediated through online spaces.

In the Web of the Parties: Local Politicians on Facebook in Hungary

The study examines the Facebook use of elected local politicians over two years in Hungary. To gain insights into the role of local politicians in social-media-based local publics in Hungary, a large-scale data collection has been conducted to capture the Facebook activity of all elected local representatives (mayors and councilors; <em>N</em> = 19,503) from the 3,152 Hungarian municipalities. Our research uncovers the level (adoption, activity) and direct audience (number of followers) of their Facebook activity and shows how these patterns are conditioned by political (party, electoral competitiveness, bandwagon effect) and contextual (size, average income of the population, development level of the local Facebook sphere) factors. We show that local politicians are mostly active in larger municipalities, while a larger proportion of the population can be reached directly in smaller communities. The activity of local politicians is largely driven by political considerations, while demand-side factors are less important.

Turning captchas against humanity: Captcha-based attacks in online social media

Nowadays, people generate and share massive amounts of content on online platforms (e.g., social networks, blogs). In 2021, the 1.9 billion daily active Facebook users posted around 150 thousand photos every minute. Content moderators constantly monitor these online platforms to prevent the spreading of inappropriate content (e.g., hate speech, nudity images). Based on deep learning (DL) advances, Automatic Content Moderators (ACM) help human moderators handle high data volume. Despite their advantages, attackers can exploit weaknesses of DL components (e.g., preprocessing, model) to affect their performance. Therefore, an attacker can leverage such techniques to spread inappropriate content by evading ACM.In this work, we analyzed 4600 potentially toxic Instagram posts, and we discovered that 44% of them adopt obfuscations that might undermine ACM. As these posts are reminiscent of captchas (i.e., not understandable by automated mechanisms), we coin this threat as Captcha Attack (CAPA). Our contributions start by proposing a CAPA taxonomy to better understand how ACM is vulnerable to obfuscation attacks. We then focus on the broad sub-category of CAPA using textual Captcha Challenges, namely CC-CAPA, and we empirically demonstrate that it evades real-world ACM (i.e., Amazon, Google, Microsoft) with 100% accuracy. Our investigation revealed that ACM failures are caused by the OCR text extraction phase. The training of OCRs to withstand such obfuscation is therefore crucial, but huge amounts of data are required. Thus, we investigate methods to identify CC-CAPA samples from large sets of data (originated by three OSN – Pinterest, Twitter, Yahoo-Flickr), and we empirically demonstrate that supervised techniques identify target styles of samples almost perfectly. Unsupervised solutions, on the other hand, represent a solid methodology for inspecting uncommon data to detect new obfuscation techniques.