4:00 PM – 4:45 PM | Merten Hall, Room 1204
Merten Hall, Room 1204
Language of deceivers: Understanding linguistic characteristics of scams on Social media platforms
Anuridhi Gupta (College of Engineering and Computing)
Social media platforms after their explosive growth, have also attracted scammers. This is why scam detection on these platforms is also becoming an ever growing challenge. Majority of the existing methods to detect scams on social media have employed machine learning and heavily focus on feature engineering. However, little is known about the explanatory characteristics of the language of these scam messages. Our study bridges this gap as we use linguistic features to find hidden clues within the messages that help distinguish between real and scam messages. In this study, we analyzed a large dataset from Twitter containing 31,000 tweets each of scam and real messages. We applied a computational linguistics approach to understand the distinct characteristics of scam and real tweet messages, in particular, by applying diversity measures at the lexical, syntactic, and semantic levels. The results show that scammers tend to use more complex sentence structures in their messages, to try to confuse the readers. These scam messages also tend to be written in a less casual manner and tend to use more words. Our extensive experimentation shows that the proposed approach of linguistic analysis is able to capture characteristics to sufficiently distinguish scam and real messages in a social media setting.
Simulated experiments: Using virtual reality for rigorous and ethical social science research
Beth Hosek (College of Education and Human Development), Stephanie Dailey
Background: Simulated experiments using human actors in virtual reality have the ability to enable researchers to explore contexts that are challenging to investigate in real life, including for reasons such as potential trauma. However, very little research emphasizes the emotional responses and reflections of operators in virtual reality using rigorous research methodology and high ethical standards. Exploring examples of simulated experiment design, implementation, and analysis implemented with rigorous research methodology and high ethical standards can justify the use of virtual reality in research contexts that might be distressing, dangerous, or potentially traumatic should they be explored in real life. Methods: Two examples of virtual reality simulated experiments will be explored: one on school shootings, and one on law enforcement use of force. Procedures of each—both individually and overall—will be discussed, to include recruitment processes, screening protocols, participant experiences in virtual reality during both training and scenarios, data collection, as well as debriefing and follow-up proceedings. Results: Participants engaged in both studies did not experience residual emotional or physical effects after fully engaging in virtual reality simulated experiments for up to two weeks. Conclusions: The two presented examples of simulated experiments using virtual reality demonstrate the safety, ethicality, and rigor possible in these contexts. Such work opens doors for future research to safely, ethically, and rigorously venture into territory not previously possible for social research.
Payment 4 Placement (P4P) at George Mason University
Gisele Hernandez (College of Public Health), Betzy Balladares
Unpaid internships can bring an abundance of hardships, and may hinder students from achieving their full potential. There is a lack of sufficient research on the hardships faced by social work students as a result of unpaid internships. Despite the professional experience gained through unpaid internships, this practice perpetuates socio-economic disparities and continues a cycle of exploitation. The prevalence of unpaid internships shows the imbalance in equity, social justice, and fairness within labor practices. The Council of Social Work Education requires Bachelor’s students to fulfill a minimum of 400 hours from their unpaid internship with supervision. Meanwhile, Master’s students from a two-year program need to to fulfill a minimum of 900 hours from their unpaid internship with supervision. Considering the hours required to obtain a degree, social work students are at risk for mental health problems, emotional distress, and financial instability.
External Validation of Computational Barthel Index: Why Accuracy Drops?
Lemba Priscille Ngana (College of Public Health), Janusz Wojtusiak
Introduction Computational Barthel Index (CBIT) is a predictive tool designed to forecast changes in patient functional status up to one year in advance, utilizing data primarily from Community Living Centers (CLCs) within the Department of Veterans Affairs (VA). Despite demonstrating high accuracy, external validation of CBIT was not conducted, potentially limiting generalizability to nursing home residents within the VA system. This study seeks to independently validate CBIT using SEER-Medicare database for both cancer and non-cancer patients. Methods SEER-Medicare data spanning from 2006 to 2018 were processed to replicate variables present in the original CBIT work. Outcome variables were derived from Minimum Data Set (MDS) 2.0 and 3.0, by applying existing methods to create indicators of activities of daily living (ADL). Datasets have been created for testing and assessing current ADLs, as well as predicting them 90, 180 and 360 days ahead. Performance was evaluated using standard metrics used in machine learning. Results The Medicare cohort consisted of 843,836 patients and 8,257,141 MDS evaluations. Descriptive analyses revealed substantial differences between the SEER-Medicare and VA datasets, particularly, differences in patient demographics and Barthel scores. More importantly, a significant decline in CBIT model accuracy was observed. Differences were also observed in performance across demographic groups. Conclusions The findings support the hypothesis that external validation of CBIT using SEER-Medicare dataset would result in decreased model performance, attributable to significant disparities between the datasets. Future research endeavors should explore the potential of transfer learning methods to mitigate these differences.
Penguin Feathers as Ecological Indicators of Regional, Global, and Historical Per- and Polyfluoroalkyl Substances (PFAS) Exposure
Matthew Badia (College of Science)
Per- and Polyfluoroalkyl Substances (PFAS) are of growing concern for their potential adverse environmental and human health effects. PFAS compounds demonstrate remarkable environmental stability and are globally dispersed. These compounds have been detected in a variety of biological samples, including bird feathers, which have proven to be a reliable indicator for the bioaccumulation of heavy metals and persistent organic pollutants. Feathers, composed mainly of keratin, accumulate PFAS, which has a documented affinity for proteins. Our study aims to determine the extent of PFAS exposure in three penguin species in the Antarctic Peninsula. Feathers of Pygoscelis adeliae (Adélie), Pygoscelis papua (Gentoo), and Aptenodytes fosteri (Emperor) were collected in 2013. Sampling sites were concentrated around the Henryk Arctouski Polish Antarctic Station, Kopaitic Island, and General Bernardo O’Higgins Base. Feathers were digested in 1.0 M potassium hydroxide in methanol for 24 hours, then neutralized with 1.0 M hydrochloric acid in methanol. Samples were passed through a 100 mg EnviCarbTM SPE tube to remove solubilized protein and analyzed for 40 legacy and emerging PFAS compounds. A Shimadzu Triple-Quadrupole Liquid Chromatography-Mass Spectrometer in Multiple Reactions Monitoring (MRM) mode was used to determine PFAS concentrations. PFAS concentrations were compared by species, age, and sampling location. Bird specimen samples have been analyzed, and preliminary results show PFAS concentrations vary between species and age group. Additionally, the proportion of precursors to carboxylic and sulfonic acids may provide insight into the source of PFAS contamination. The results of this study have significance in determining the pervasiveness of PFAS in the Antarctic Peninsula.
In Tune with Change: A Mixed Methods Study of Reconstructionist Philosophy and the Transformative Power of Music of Diverse Cultures
MeHaley Babich (College of Visual and Performing Arts)
As music education strives to be a transformative force in society, this mixed methods study examines the experiences, beliefs, and attitudes of music educators concerning the intersection of the use of multicultural music and guiding objectives of the Reconstructionist Philosophy of Music Education (RPME). The need for this study is to highlight how and why teachers incorporate social change practices in their curricula, through multicultural music and RPME objectives, despite having no formal education or requirement to adhere to guiding objectives. This research aims to uncover the pivotal role of RPME in promoting multicultural music for social change in the music classroom. The questions used to guide this research are: 1. What are the experiences, beliefs, and attitudes of music educators regarding the guiding objectives of RPME and their role in promoting multicultural music for social change in classrooms? and 2. To what degree do music educators incorporate multicultural music for social change in their curricula? The design of this study will be a descriptive exploratory sequential mixed methods design. The qualitative strand will explore music educators’ experiences, beliefs and attitudes, while the quantitative strand will measure integration and assess familiarity with RPME objectives. By understanding this phenomenon and frequency of use, the outcomes of this study could provide advocacy that cements the need for music education classes in K – 12 education, advocacy for professional development in RPME, and recommendations for changes to current curricula practices for educators, policymakers, and researchers interested in enhancing the transformative potential of music education.
Exploring the Trajectory of PTSD Symptoms and Pain Interference in HIV+ Women
Patricia Mejia M.S. (College of Humanities and Social Sciences), Leah Adams Ph.D.
In recent years, anti-retroviral therapy has greatly increased the quality of life for persons living with Human Immunodeficiency Virus (HIV; Kietrys et al., 2020). However, despite advances in medical care, pain remains a common comorbidity among individuals living with HIV (Crocket et al., 2020). Further, less is known about psychological factors, such as Posttraumatic Stress Disorder (PTSD), that may impact the experience of pain (Ayano et al., 2020). While relevant, few studies have examined the relation between pain-related interference and PTSD symptoms in tandem (Bhatraju et al., 2023), and even less is known of how they evolve over time. The purpose of this study was to examine the trajectories of PTSD symptomatology and pain interference over time in a sample of HIV+ women living with pain. A total of 415 HIV+ women living with pain completed self-report measures during their routine medical follow-ups from four time points (i.e., 2012-2015). A bivariate latent growth curve model was employed to evaluate the trajectories of pain interference and PTSD symptoms concurrently over the specified time points. Findings indicated that pain interference significantly increased in HIV+ women over time. While PTSD symptoms did increase, they did not increase significantly over time. The findings underscore the need for targeted interventions to address the burden of pain interference in this population while monitoring PTSD symptoms that appear stable but are entwined with pain.
Reducing the fear and pain during vaccinations in children
Patricia Solis (College of Science)
Type of submission: Quality Improvement Learning Objectives Goal #1:Decrease fear and pain experienced by children during vaccinations by teaching the vaccination team and staff available techniques. Goal #2: Reducing vaccination avoidance in children by implementing coping techniques by the vaccination team and staff. Goal #3: Increase efficiency and disseminate toolkit recommendations to be used in current and other vaccination clinics. Background/ Purpose Pain associated with vaccinations is a source of distress for children, their parents, and those administering the vaccinations. Vaccinations are the most common painful needle procedure experienced throughout life. According to the CDC vaccination recommendations, about 30 vaccinations are given to a child before the child’s second birthday. Among children, 63%, and adolescents, 50% fear vaccinations, and if left untreated, this fear leads to 25% of vaccination avoidance in adults. This quality improvement paper presents a proposal for solutions to fear and pain in children during vaccinations at immunization clinics by introducing the Fear and Pain Management Strategies Toolkit. Methods: The methodology in this project uses the Wong-Baker pain scale and questions addressed in survey form to children, parents, and vaccination staff about the fear and pain experienced during vaccinations. Four surveys ask parents and children about preferred coping techniques, pain and fear level before vaccinations, pain and fear level after vaccinations, and the likelihood of adhering to vaccinations using coping techniques. Findings/Results (in Progress): The supporting evidence results in positive outcomes when implementing coping techniques for fear and pain of vaccination in children. Educating children, parents, and staff on coping interventions by using these techniques has led to positive experiences about immunization pain and fear. Implications for Nursing Practice: Educating the vaccination team on these coping techniques can reduce vaccine avoidance in children as they grow into adults, increase parental commitment to regular immunization, and enhance the overall efficiency of the vaccination clinic.
Sex Workers Experiences of Enacted Stigma Reported During Podcast Interviews: A Qualitative Case Study
Tara Altay (College of Public Health)
Introduction/Purpose Stigma’s negative effect on sex workers (SW) health has been studied among various populations of SW. However, research exploring enacted stigma (ES), the behaviors attitudes and statements of providers toward SW, in healthcare encounters in the US is understudied. The purpose of this study was to explore SW experiences of ES in healthcare encounters in the United States. Research Questions 1. What are SW experiences of ES during healthcare encounters and how are SW describing ES among themselves without non-SW influence during interviews? 2. How are SW navigating ES in healthcare encounters? Methods This case study design, with 3 embedded units of analysis, explored podcast episodes/archives hosted by SW who interviewed SW and discussed ES in healthcare encounters (Stake, 1995). Podcast episode analysis was conducted through a critical, etic lens employing grounded theory techniques. Podcast archives were explored according to Lundstrom and Lundstrom (2020) method to locate the podcast’s mission and identify political ideologies of the hosts, thus further aligning with critical epistemology (Aranda, 2020). Results Four main themes were constructed from the data including: provider assumptions, false caricature of a SW, sexually transmitted infections and refer to nontraditional healthcare providers. Discussion/Conclusion Providers can relinquish stereotypes and binaries of good/bad people or choices. Providers can make certain decisions personally in terms of what they consider deviant/healthy but in patient encounters providers need to center patient opinions and believe SW appraisals of their lives while offering expertise based on presenting symptomatology and in response to patient requests and concerns.
Early Black Librarianship and Library Services
Timmia King (College of Humanities and Social Sciences)
This poster proposes to explore a history of Black librarianship in the United from 1899-1942 using a dataset called the Directory of Negro Graduate of Accredited Library schools 1900-1936, data in early Journal articles such as the “Negro Library Workers” (1940) by Wallace Van Jackson and the “Library Service and the Negro” (1932) by Louis Shores, and census data. These resources contain a wealth of data. The directory was published by the Columbia Civic Library Association, a professional organization for Black librarians and library workers that was based in the District of Columbia. It included a list of graduates, the institution and the year they graduated, and their current work location. These datasets and supplementary sources, although incomplete, taken together better help to illuminate an early picture who these Black Librarians were, the library training available to them, and the geographic areas and types of libraries that they worked. Visualizing the dataset in different ways such as map data visualizations allows for us to see types of patterns or clusters and the supplementary information helps to explain why these patterns existed. Libraries are places for educations. These early Black librarians’ work was important to the communities that they served, they curated and created knowledge and then made it available. These places for education were opened up to different segments of the population. This poster will provide a look into that process over time showing relationship between the growth of Black Librarianship and Library Services to African Americans.
Predicting Scam Detection Performance on Social Media using Eye Tracking
Tomas Lapnas (College of Humanities and Social Sciences)
Cybercrime has been on the rise in recent years, posing a significant threat to users online. Phishing attacks, which utilize social engineering techniques to extract sensitive information from users, are the most common form of cybercrime. To understand how users assess the legitimacy of online content, researchers use eye tracking to directly measure gaze patterns and predict behavior. However, there is a lack of literature investigating phishing on social media platforms using such technologies. In our study, 18 participants read 101 tweets and provided dichotomous responses indicating the presence or absence of malicious content. Participants were sensitive to the presence of the scams (d’ = 0.87), especially when tweets contained financial information (d’ = 1.49; t(31.32) = -3.10, p = 0.004). Eye tracking metrics proved to be strongly correlated to overall scam detection performance. Regressive saccades, a type of eye movement indicating reanalysis of prior information, in conjunction with stimulus type (legitimate, scam), accounted for 49% of overall performance variance. On trials only pertaining to scams, regressive saccades alone were significantly negatively correlated with performance (r = -0.86), accounting for 74% of performance variance. Using LIWC, a state-of-the-art text analysis software, we were about to identify linguistic differences between legitimate and scam tweets. Scams were significantly higher in positive emotion and tone, sentiment pertaining to achievement, and references to the future. By understanding how users visually inspect information online, we are better able to protect users and identify potential threats to cybersecurity.
Social ties in systemic crises: The role of non-coresident kin in intercity migration dynamics during COVID-19
Unchitta Kan (College of Science), Jericho McLeod, Eduardo López
Changes in U.S. migration during the COVID-19 pandemic indicate that many moved to less populated cities from larger cities, deviating from previous trends. Considering prior work in the literature showing the distribution of kin ties in the U.S., in which people are more likely to have kin ties in smaller cities, we ask a question that has been overlooked: is this trend in pandemic-migration partially a consequence of individuals relocating to be closer to kin? Employing two large-scale data sets, census microdata and mobile phone GPS relocation data, we analyze the three-way relation between intercity migration, family ties, and city population size. We perform our analyses both at the individual level and at the city level. Our collection of empirical results paints a picture of migration change affected by kin. Namely, we find that people migrated closer to parental family at higher rates after the COVID-19 pandemic started. Moreover, even controlling for factors such as population density and cost of living, changes in net in-migration tended to be larger and positive in cities with larger proportions of people who can be parents to adult children (our proxy for parental family availability, which is also inversely related to population size). Our study advances the emerging demography-disaster nexus and amplifies ongoing literature highlighting the role of broader kinship systems in large-scale socioeconomic phenomena.
Development of PM2.5 transport: Modeling the spatial distribution of Camp Fire from California to New York
Xiaorong Shan (College of Engineering and Computing), Daniel Tong, Joan Casey, Lucas Henneman
The November 2018 Camp Fire that took place in northern California, releasing abundant of aerosols into the atmosphere, one of the costliest disasters in the world for insurers with insured losses totaling $12.5 billion. The event degraded air quality across the United States, with elevated PM2.5 concentrations observed along the East Coast during this period. It is unclear, however, the extent that elevated air pollution concentrations attributable to distant wildfire emissions impacted human health. Modeled wildfire source impacts are highly dependent on the models used and their inputs, including wildfire emissions, which can be highly uncertain. These differences may propagate or cause under or over estimations over large areas in public health analysis. Here we focus on PM2.5 exposures during the Camp Fire period (from Nov. 08.2018 to Dec. 02.2018) in New York state. To predict the exposure, we use the average values of four satellite-based fire emission data sets (FEER, FLAMBE, GBBEPx, and GFAS). Then, we employ HyADS, the HYSPLIT average dispersion model, which combines the HYSPLIT trajectory dispersion with emissions to simulate the PM2.5 daily exposure. Exposure was low at the beginning of the Camp Fire period, but it increased drastically from Nov.24 to Nov. 26. We evaluate our results PM2.5 observations at EPA monitors, which capture total daily variability but are limited in their ability to differentiate PM2.5 from fires, to compare the variation trend of these two group results. The highest correlation (R) between the HyADS and local monitors is 0.54 when the most detailed emissions data. In continuing work, we plan to employ source apportionment at EPA CSN cites to identify the portion of the observed PM attributable to fires. Also, we will leverage CMAQ model to allow more emission and meteorological factors to be considered in PM2.5 transportation.