Olga Scrivner has a MCS from UIUC and a dual PhD from IU. Her main interest is machine learning, CL, NLP, data mining, visualization, and deep learning to unveil data patterns and build actionable applications. Olga is an Assistant Professor in Computer Science and Software Engineering at Rose-Hulman Institute of Technology. She also teaches Data Analytics courses at Harrisburg University of Science and Technology and she is an Adjunct Professor in Data Science program at Indiana University. She has also developed two Shiny web applications for text mining and quantitative analysis: ITMS and LVS. Currently, she also explores 1) the use of emerging technologies (virtual and augmented reality) in teaching and learning; 2) learning analytics; 3) the development of web data visualization applications; 4) labor market-education skill gap analysis; 5) academic ROI via funding and publications; 6) effect of opioid epidemic on job market. Supervisors: Barbara Vance, Sandra Kübler, Julie Auger, Markus Dickinson, and Marco Passarotti
Presentation at the Complexity and Cognition CCS 2020 Conference.
Psychosocial stress is a commo... more Presentation at the Complexity and Cognition CCS 2020 Conference.
Psychosocial stress is a common factor contributing to the increase in opioid misuse. Stressors as socio-environmental demands on individuals (e.g., trauma, adverse childhood events, unemployment, stigma, food insecurity) have been shown to also impact cognitive processes. Particularly, the impact has been noted on executive functioning (EF) linked to reasoning, problem-solving, decision-making. Understanding the relationship between interpersonal trauma, opioid use disorder (OUD) and cognitive impairments is essential for the rehabilitation trajectory and consequent improvement of social life.We assess 10 psychosocial stressors: social stability, food insecurity, substance use and severity, drug use stigma, social support, perceived stress, depression, anxiety, adverse childhood experiences, and trauma. Executive functions are measured via two computer-assisted tests: Iowa Gambling Task, and Opioid Word Stroop Test.
2022 IEEE Global Engineering Education Conference (EDUCON)
It is known that timely and personalized feedback is vital to the learning process, and because o... more It is known that timely and personalized feedback is vital to the learning process, and because of increasing enrollment, instructors can find it harder to provide that feedback. Learning analytics presents a solution to this problem. The growth in popularity of online education systems better enables learning analytics by providing additional educational data. This work focuses on the analysis of students' incorrect short answers and their pathways to correct solutions. By considering student submissions as sequences, this work uses a dimension called "distance" which can be used to predict how far off a student's incorrect answer is from a correct one. This distance metric can be used for recognizing students who may need help, understanding which concepts students struggle with, evaluating assessment questions, and improving multiple-choice answers. This paper discusses the methods, relevant learning scenarios, and applications of the learning analytics system. It features the results and analysis of a usability test conducted on 56 faculty members.
Access to a large amount of scholarly publication presents new opportunities to researchers. Rece... more Access to a large amount of scholarly publication presents new opportunities to researchers. Recent advances in data visualization techniques allow for automated content analysis, topic modeling and classification as well as research trend and scientific network analyses. Many probabilistic models and text-mining tools have been put forth to help analyze and explore large collections of documents. The use of these tools in mainstream scholarly research, however, remains limited to machine learning and natural language processing fields, as the researcher is often challenged by the technological hurdle of command-line tools. We address this issue by introducing a user-friendly application that allows researchers to visually explore scholarly articles. Written in R with the Shiny web app, this application not only provides a web-interactive interface, but also allows researchers to implement state-of-the-art topic modeling and visualization tools. Finally, the accessibility of our web...
Aim: Psychosocial stress is known to affect substance use disorders. Little is known about how ps... more Aim: Psychosocial stress is known to affect substance use disorders. Little is known about how psychosocial stress relates to cognitive function and trajectories of opioid use disorder (OUD), specifically. Executive function (EF) is known to be impaired in substance abusers, and poor EF may adversely impact treatment adherence and treatment for OUD. The aim of this pilot study was to explore psychosocial factors associated with EF. Methods: Community-based recruitment occurred using advertisements and referral from study participants. Eligibility included 18 years or older, opioid use in the past 90 days, and screening positive for DSM-5 OUD. Participants completed a questionnaire about substance use, social instability, food insecurity, drug use stigma, social support, perceived stress, depression, anxiety, and trauma. Executive function was assessed using a computerized version of the Iowa Gambling Test, a measure of risky decision-making, such that a higher proportion of advantageous selections (low risk, low reward) indicated better EF. Robust linear regression adjusting for age, race, education, and opioid severity was used to identify psychosocial factors associated with EF. Results: The 46 participants were primarily male, white non-Hispanic, and >40 years; 89% used opioids at least weekly, 74% used within 48 hours of study participation, and polysubstance use was common. Higher anticipated drug use stigma was associated with worse EF (B= -3.05, p=0.01); higher emotional social support was associated with better EF (B=4.10, p=0.01). Higher food insecurity (B=0.43, p=0.01) and moderate cannabis use (B=6.04, p=0.03) were also associated with better EF. Conclusion: Higher social support, lower stigma, and, paradoxically, more food insecurity and cannabis use were all strongly associated with better EF in individuals with OUD. Interventions that focus on social support and stigma could support improvements in EF and may improve OUD treatment retention (behavioral and pharmacological). Further studies are needed to clarify paradoxical findings.
In the last few decades there has been a proliferation of institutional initiatives to promote fa... more In the last few decades there has been a proliferation of institutional initiatives to promote faculty excellence and innovation in teaching and learning. Among the top-down and bottom-up approaches, the teaching innovation is effectively shown with “a participatory, collaborative methods to identify problems and solutions” and sharing leadership among all stakeholders. From an organizational perspective, faculty development has been conceptualized as a taxonomy with three levels of engagement: good teaching, scholarly teaching, and the scholarship of teaching and learning. Similar to a community-of-practice, the members of Scholarship of Teaching and Learning (SoTL) are driven by a shared interest and enthusiasm to improve their teaching and learning. Representing time-varying SoTL events and relationships between SoTL members as a community network introduces challenges in data linking, data model, and network analysis. In particular, it is essential to design solutions to preserve the network topology, temporal information, member status transformation, and diverse relationships between nodes. In order to account for the SoTL network complexity, we design a heterogenous graph model and perform co-authorship and event network analyses to evaluate the effectiveness of the current SoTL strategies in attracting new members and supporting the sustainability of existing cohorts and provide data-driven decision support for SoTL programs in their development and priorities.
Effective treatment strategies exist for substance use disorder (SUD), however severe hurdles rem... more Effective treatment strategies exist for substance use disorder (SUD), however severe hurdles remain in ensuring adequacy of the SUD treatment (SUDT) workforce as well as improving SUDT affordability, access and stigma. Although evidence shows recent increases in SUD medication access from expanding Medicaid availability under the Affordable Care Act, it is yet unknown whether these policies also led to a growth in the changes in the nature of hiring in SUDT related workforce, partly due to poor data availability. Our study uses novel data to shed light on recent trends in a fast-evolving and policy-relevant labor market, and contributes to understanding the current SUDT related workforce and the effect of Medicaid expansion on hiring attempts in this sector. We examine attempts over 2010-2018 at hiring in the SUDT and related behavioral health sector as background for estimating the causal effect of the 2014-and-beyond state Medicaid expansion on these outcomes through "differ...
In recent years, there has been growing interest in data visualization for text analysis. While t... more In recent years, there has been growing interest in data visualization for text analysis. While text mining and visualization tools have evolved into mainstream research methods in many fields (e.g. social sciences, machine learning), their application to linguistic and literary studies still remains infrequent. First, many text processing tools require some programming skills, which take time to learn and are often challenging for digital humanities scholars. Secondly, while some visualization tools (e.g. Voyant, Weka and PaperMachine) provide graphical-user interfaces, social and humanities researchers seek more interactive and dynamic control of tools, which can serve as "holistic support for exploratory analysis" (Klein 2013). This workshop introduces two user-friendly applications, namely Language Variation Suite and Interactive Text Mining Suite, that allow researchers visually explore and statistically analyze language data. Written in R with Shiny app, these applications not only provide a web interactive interface, they also allow researchers implement state-of-the-art statistical methods, such as cluster analysis, topic modeling, inferential trees and mixed model logistic regressions.
The Complex Networks and Systems NSF Research Traineeship (CNS-NRT) is an interdisciplinary multi... more The Complex Networks and Systems NSF Research Traineeship (CNS-NRT) is an interdisciplinary multifaceted STEM program with a rigorous dual-PhD degree focus, emphasizing collaborative skills and workforce development. This complex institutional training model requires comprehensive evolutionary systems thinking and modeling, and novel data visualizations to communicate key results to diverse stakeholders. The Systems Evaluation Protocol (SEP) enables the inclusion of multiple perspectives, reflecting the complexity of program activities and outcomes. Through the SEP cyberinfrastructure platform, the Netway, the CNS NRT evaluation is visually modeled by logically connecting all program activities with short-, mid-, and long-term outcomes, building a comprehensive set of networked pathways that form the basis for an evaluation plan that integrates institutional data metrics and other measurement systems and enables comparison of CNS-NRT student outcomes with those of single PhD CNS and non-NRT dual PhD students. This paper describes the pathway model visualizations and the resulting comprehensive evaluation plan.
Understanding the emergence, co-evolution, and convergence of science and technology (S&T) areas ... more Understanding the emergence, co-evolution, and convergence of science and technology (S&T) areas offers competitive intelligence for researchers, managers, policy makers, and others. This paper presents new funding, publication, and scholarly network metrics and visualizations that were validated via expert surveys. The metrics and visualizations exemplify the emergence and convergence of three areas of strategic interest: artificial intelligence (AI), robotics, and internet of things (IoT) over the last 20 years (1998-2017). For 32,716 publications and 4,497 NSF awards, we identify their topical coverage (using the UCSD map of science), evolving co-author networks, and increasing convergence. The results support data-driven decision making when setting proper research and development (R&D) priorities; developing future S&T investment strategies; or performing effective research program assessment.
Teaching Language and Teaching Literature in Virtual Environments, 2018
Recent advances in technology have made it possible to add immersive interactive dimensions to ne... more Recent advances in technology have made it possible to add immersive interactive dimensions to nearly any learning environment. This immersive technology provides students with active control and more authentic experiences; thus, helping them learn more effectively and increase their retention. In this view, these technologies seem to be an ‘ideal’ instrument for language instruction, as they combine visual, auditory, and kinesthetic learning styles. While their digital capabilities are almost limitless, their use in language learning remains limited due to technological and methodological challenges. This chapter provides methodological recommendations for the design and use of augmented and virtual technologies in language classroom settings. At the core of these recommendations is the collaborative research conducted at Indiana University which investigated the impact of mobile immersive technology for foreign language teaching and learning. Based on the findings, this chapter suggests several immersive tools and applications suited for the use in foreign language classroom (Aurasma, ThingLink, and Google Cardboard), which were evaluated by both students and instructors by means of self-assessment, technical feedback, and usage statistics.
Presentation at the Complexity and Cognition CCS 2020 Conference.
Psychosocial stress is a commo... more Presentation at the Complexity and Cognition CCS 2020 Conference.
Psychosocial stress is a common factor contributing to the increase in opioid misuse. Stressors as socio-environmental demands on individuals (e.g., trauma, adverse childhood events, unemployment, stigma, food insecurity) have been shown to also impact cognitive processes. Particularly, the impact has been noted on executive functioning (EF) linked to reasoning, problem-solving, decision-making. Understanding the relationship between interpersonal trauma, opioid use disorder (OUD) and cognitive impairments is essential for the rehabilitation trajectory and consequent improvement of social life.We assess 10 psychosocial stressors: social stability, food insecurity, substance use and severity, drug use stigma, social support, perceived stress, depression, anxiety, adverse childhood experiences, and trauma. Executive functions are measured via two computer-assisted tests: Iowa Gambling Task, and Opioid Word Stroop Test.
2022 IEEE Global Engineering Education Conference (EDUCON)
It is known that timely and personalized feedback is vital to the learning process, and because o... more It is known that timely and personalized feedback is vital to the learning process, and because of increasing enrollment, instructors can find it harder to provide that feedback. Learning analytics presents a solution to this problem. The growth in popularity of online education systems better enables learning analytics by providing additional educational data. This work focuses on the analysis of students' incorrect short answers and their pathways to correct solutions. By considering student submissions as sequences, this work uses a dimension called "distance" which can be used to predict how far off a student's incorrect answer is from a correct one. This distance metric can be used for recognizing students who may need help, understanding which concepts students struggle with, evaluating assessment questions, and improving multiple-choice answers. This paper discusses the methods, relevant learning scenarios, and applications of the learning analytics system. It features the results and analysis of a usability test conducted on 56 faculty members.
Access to a large amount of scholarly publication presents new opportunities to researchers. Rece... more Access to a large amount of scholarly publication presents new opportunities to researchers. Recent advances in data visualization techniques allow for automated content analysis, topic modeling and classification as well as research trend and scientific network analyses. Many probabilistic models and text-mining tools have been put forth to help analyze and explore large collections of documents. The use of these tools in mainstream scholarly research, however, remains limited to machine learning and natural language processing fields, as the researcher is often challenged by the technological hurdle of command-line tools. We address this issue by introducing a user-friendly application that allows researchers to visually explore scholarly articles. Written in R with the Shiny web app, this application not only provides a web-interactive interface, but also allows researchers to implement state-of-the-art topic modeling and visualization tools. Finally, the accessibility of our web...
Aim: Psychosocial stress is known to affect substance use disorders. Little is known about how ps... more Aim: Psychosocial stress is known to affect substance use disorders. Little is known about how psychosocial stress relates to cognitive function and trajectories of opioid use disorder (OUD), specifically. Executive function (EF) is known to be impaired in substance abusers, and poor EF may adversely impact treatment adherence and treatment for OUD. The aim of this pilot study was to explore psychosocial factors associated with EF. Methods: Community-based recruitment occurred using advertisements and referral from study participants. Eligibility included 18 years or older, opioid use in the past 90 days, and screening positive for DSM-5 OUD. Participants completed a questionnaire about substance use, social instability, food insecurity, drug use stigma, social support, perceived stress, depression, anxiety, and trauma. Executive function was assessed using a computerized version of the Iowa Gambling Test, a measure of risky decision-making, such that a higher proportion of advantageous selections (low risk, low reward) indicated better EF. Robust linear regression adjusting for age, race, education, and opioid severity was used to identify psychosocial factors associated with EF. Results: The 46 participants were primarily male, white non-Hispanic, and >40 years; 89% used opioids at least weekly, 74% used within 48 hours of study participation, and polysubstance use was common. Higher anticipated drug use stigma was associated with worse EF (B= -3.05, p=0.01); higher emotional social support was associated with better EF (B=4.10, p=0.01). Higher food insecurity (B=0.43, p=0.01) and moderate cannabis use (B=6.04, p=0.03) were also associated with better EF. Conclusion: Higher social support, lower stigma, and, paradoxically, more food insecurity and cannabis use were all strongly associated with better EF in individuals with OUD. Interventions that focus on social support and stigma could support improvements in EF and may improve OUD treatment retention (behavioral and pharmacological). Further studies are needed to clarify paradoxical findings.
In the last few decades there has been a proliferation of institutional initiatives to promote fa... more In the last few decades there has been a proliferation of institutional initiatives to promote faculty excellence and innovation in teaching and learning. Among the top-down and bottom-up approaches, the teaching innovation is effectively shown with “a participatory, collaborative methods to identify problems and solutions” and sharing leadership among all stakeholders. From an organizational perspective, faculty development has been conceptualized as a taxonomy with three levels of engagement: good teaching, scholarly teaching, and the scholarship of teaching and learning. Similar to a community-of-practice, the members of Scholarship of Teaching and Learning (SoTL) are driven by a shared interest and enthusiasm to improve their teaching and learning. Representing time-varying SoTL events and relationships between SoTL members as a community network introduces challenges in data linking, data model, and network analysis. In particular, it is essential to design solutions to preserve the network topology, temporal information, member status transformation, and diverse relationships between nodes. In order to account for the SoTL network complexity, we design a heterogenous graph model and perform co-authorship and event network analyses to evaluate the effectiveness of the current SoTL strategies in attracting new members and supporting the sustainability of existing cohorts and provide data-driven decision support for SoTL programs in their development and priorities.
Effective treatment strategies exist for substance use disorder (SUD), however severe hurdles rem... more Effective treatment strategies exist for substance use disorder (SUD), however severe hurdles remain in ensuring adequacy of the SUD treatment (SUDT) workforce as well as improving SUDT affordability, access and stigma. Although evidence shows recent increases in SUD medication access from expanding Medicaid availability under the Affordable Care Act, it is yet unknown whether these policies also led to a growth in the changes in the nature of hiring in SUDT related workforce, partly due to poor data availability. Our study uses novel data to shed light on recent trends in a fast-evolving and policy-relevant labor market, and contributes to understanding the current SUDT related workforce and the effect of Medicaid expansion on hiring attempts in this sector. We examine attempts over 2010-2018 at hiring in the SUDT and related behavioral health sector as background for estimating the causal effect of the 2014-and-beyond state Medicaid expansion on these outcomes through "differ...
In recent years, there has been growing interest in data visualization for text analysis. While t... more In recent years, there has been growing interest in data visualization for text analysis. While text mining and visualization tools have evolved into mainstream research methods in many fields (e.g. social sciences, machine learning), their application to linguistic and literary studies still remains infrequent. First, many text processing tools require some programming skills, which take time to learn and are often challenging for digital humanities scholars. Secondly, while some visualization tools (e.g. Voyant, Weka and PaperMachine) provide graphical-user interfaces, social and humanities researchers seek more interactive and dynamic control of tools, which can serve as "holistic support for exploratory analysis" (Klein 2013). This workshop introduces two user-friendly applications, namely Language Variation Suite and Interactive Text Mining Suite, that allow researchers visually explore and statistically analyze language data. Written in R with Shiny app, these applications not only provide a web interactive interface, they also allow researchers implement state-of-the-art statistical methods, such as cluster analysis, topic modeling, inferential trees and mixed model logistic regressions.
The Complex Networks and Systems NSF Research Traineeship (CNS-NRT) is an interdisciplinary multi... more The Complex Networks and Systems NSF Research Traineeship (CNS-NRT) is an interdisciplinary multifaceted STEM program with a rigorous dual-PhD degree focus, emphasizing collaborative skills and workforce development. This complex institutional training model requires comprehensive evolutionary systems thinking and modeling, and novel data visualizations to communicate key results to diverse stakeholders. The Systems Evaluation Protocol (SEP) enables the inclusion of multiple perspectives, reflecting the complexity of program activities and outcomes. Through the SEP cyberinfrastructure platform, the Netway, the CNS NRT evaluation is visually modeled by logically connecting all program activities with short-, mid-, and long-term outcomes, building a comprehensive set of networked pathways that form the basis for an evaluation plan that integrates institutional data metrics and other measurement systems and enables comparison of CNS-NRT student outcomes with those of single PhD CNS and non-NRT dual PhD students. This paper describes the pathway model visualizations and the resulting comprehensive evaluation plan.
Understanding the emergence, co-evolution, and convergence of science and technology (S&T) areas ... more Understanding the emergence, co-evolution, and convergence of science and technology (S&T) areas offers competitive intelligence for researchers, managers, policy makers, and others. This paper presents new funding, publication, and scholarly network metrics and visualizations that were validated via expert surveys. The metrics and visualizations exemplify the emergence and convergence of three areas of strategic interest: artificial intelligence (AI), robotics, and internet of things (IoT) over the last 20 years (1998-2017). For 32,716 publications and 4,497 NSF awards, we identify their topical coverage (using the UCSD map of science), evolving co-author networks, and increasing convergence. The results support data-driven decision making when setting proper research and development (R&D) priorities; developing future S&T investment strategies; or performing effective research program assessment.
Teaching Language and Teaching Literature in Virtual Environments, 2018
Recent advances in technology have made it possible to add immersive interactive dimensions to ne... more Recent advances in technology have made it possible to add immersive interactive dimensions to nearly any learning environment. This immersive technology provides students with active control and more authentic experiences; thus, helping them learn more effectively and increase their retention. In this view, these technologies seem to be an ‘ideal’ instrument for language instruction, as they combine visual, auditory, and kinesthetic learning styles. While their digital capabilities are almost limitless, their use in language learning remains limited due to technological and methodological challenges. This chapter provides methodological recommendations for the design and use of augmented and virtual technologies in language classroom settings. At the core of these recommendations is the collaborative research conducted at Indiana University which investigated the impact of mobile immersive technology for foreign language teaching and learning. Based on the findings, this chapter suggests several immersive tools and applications suited for the use in foreign language classroom (Aurasma, ThingLink, and Google Cardboard), which were evaluated by both students and instructors by means of self-assessment, technical feedback, and usage statistics.
Proceedings of the National Academy of Sciences, 2018
Rapid research progress in science and technology (S&T) and continuously shifting workforce needs... more Rapid research progress in science and technology (S&T) and continuously shifting workforce needs exert pressure on each other and on the educational and training systems that link them. Higher education institutions aim to equip new generations of students with skills and expertise relevant to workforce participation for decades to come, but their offerings sometimes misalign with commercial needs and new techniques forged at the frontiers of research. Here, we analyze and visualize the dynamic skill (mis-)alignment between academic push, industry pull, and educational offerings, paying special attention to the rapidly emerging areas of data science and data engineering (DS/DE). The visualizations and computational models presented here can help key decision makers understand the evolving structure of skills so that they can craft educational programs that serve workforce needs. Our study uses millions of publications, course syllabi, and job advertisements published between 2010 a...
Proceedings of the Midwest Instruction and Computing Symposium 2024, 2024
Artificial Intelligence (AI) holds immense promise in revolutionizing healthcare, offering potent... more Artificial Intelligence (AI) holds immense promise in revolutionizing healthcare, offering potential benefits for patients and healthcare professionals. However, its full potential remains constrained by significant barriers to adoption within the healthcare industry. This research investigates the regulatory, operational, and ethical domains surrounding AI software in healthcare, more specifically in the field of Radiology. The main objectives are to 1) examine the current barriers to AI implementation from different perspectives: manufacturer, radiologist, technologist, and sales representative, 2) evaluate the current regulation taxonomies (AI EU Act and FDA AI approval procedures), and 3) propose an extended framework increasing the practicality of current AI regulations. The research methodology involved a qualitative survey administered at the Annual Radiological Society Conference. Survey respondents provided valuable insights into existing software solutions and adoption barriers, informing the extension of current models for AI regulations. The preliminary findings suggested a five-pillar classification system for radiological AI-enabled devices: learning, documenting, planning, communication, and discovery applications. This classification system offers a more nuanced and accurate representation as compared to the FDA's existing three-class system, based on the risk and impact on the patient, from low to high-risk applications. The findings also revealed the potential "safety" gap in the current FDA regulation.
In the last few decades there has been a proliferation of institutional initiatives to promote f... more In the last few decades there has been a proliferation of institutional initiatives to promote faculty excellence and innovation in teaching and learning. Among the top-down and bottom-up approaches, the teaching innovation is effectively shown with “a participatory, collaborative methods to identify problems and solutions” and sharing leadership among all stakeholders. From an organizational perspective, faculty development has been conceptualized as a taxonomy with three levels of engagement: good teaching, scholarly teaching, and the scholarship of teaching and learning. Similar to a community-of-practice, the members of Scholarship of Teaching and Learning (SoTL) are driven by a shared interest and enthusiasm to improve their teaching and learning. Representing time-varying SoTL events and relationships between SoTL members as a community network introduces challenges in data linking, data model, and network analysis. In particular, it is essential to design solutions to preserve the network topology, temporal information, member status transformation, and diverse relationships between nodes. In order to account for the SoTL network complexity, we design a heterogenous graph model and perform co-authorship and event network analyses to evaluate the effectiveness of the current SoTL strategies in attracting new members and supporting the sustainability of existing cohorts and provide data-driven decision support for SoTL programs in their development and priorities.
Shifting from a teacher-centered to a student-centered education we are inevitably entering a puz... more Shifting from a teacher-centered to a student-centered education we are inevitably entering a puzzling world of digital natives who “live and breathe on social media”. Despite our efforts to learn and adapt new technology, the gaps between generations X, Y and Z keep growing. By the time we, digital immigrants, have incorporated Facebook into our classrooms, digital natives have already moved to other visually engaging platforms, e.g. Whisper and Snapchat. However, the digital bridge between our generations is not simply in using new technology in the classroom. The key element is in the understanding of social media and choosing the right social channel for you and your students. For instance, traditional e- mail communications with our students can be improved by using Tweeter feeds. Edmodo can further help us creating a social digital classroom and our digital course materials can be organized using Pinterest. Finally, why not let generation Z create their projects in their digital home using Tumblr and Storify? This workshop aims to provide a review of social media channels for classroom use. I will describe the cons and pros for most commonly used media and will discuss how to choose the right ones depending on the classroom goals.
In recent years there has been growing interest in quantitative methods for analyzing linguistic ... more In recent years there has been growing interest in quantitative methods for analyzing linguistic data. Advanced multifactorial statistical analyses, such as inferential trees, mixed-effect logistic regression and Bayesian hierarchical models, have become more accessible for linguistic research as a result of the availability of an open source programming environment provided by the statistical software R. In the present paper, we introduce a novel toolkit, Language Variation Suite, a software program that comprises a friendly environment for conducting quantitative analyses. We demonstrate how the theory built on traditional mono-factorial analysis can be extended to a multifactorial approach allowing for a deeper understanding of language variation. The focus of the analysis is based on intervocalic /d/ deletion in Spanish. Various factors have been proposed to account for this phenomenon (e.g., stress Hualde et al. 2011, syntactic category Toreblanca 1976, stylistics and geographical location Quilis 1993, Lewis 2001). For our investigation we have selected thirty two speaker from the Diachronic Study of the Speech of Caracas 1987 and 2004-2010. In contrast to traditional methodological approaches we have treated intervocalic /d/ as a continuous dependent variable according to the intensity ratio measurements obtained. This measurement was obtained by dividing the lowest intensity point of /d/ by the highest intensity point of the preceding vowel. This equation provided a value between 1, a more vowel like production, and 0, a more stop like production (Carrasco, Hualde & Simonet, 2012). Furthermore, we have integrated various syntactic, phonetic and sociolinguistic factors. First, the macro view of the data is performed: the most significant variables for each category are identified by using a Random Tree analysis. These variables are further fitted into regression models, which provide an overall picture of our data: intervocalic /d/ tends to be produced more as a vowel by younger male speaker in the context of a following low/mid vowel within the syntactic category of participle and frequent word types. These findings are further confirmed by a robust mixed regression model. Second, the micro view is performed by means of inferential trees. The results show that young and middle age male speakers tend to produce more lenited variants of /d/ in 1987, whereas in 2007 middle age speakers produce less lenited variants, similar to older speaker. On phonetic level, more lenited /d/ occurs with following low/mid vowels that are not part of diphthongized sequence. The present analysis contributes to demonstrate the implementation of a multifactorial macro and micro analysis of /d/ lenition, which can be constructed in a dynamic environment by using a novel toolkit: Language Variation Suite.
The present investigation contributes to further our knowledge on the pattern of variation involv... more The present investigation contributes to further our knowledge on the pattern of variation involving intervocalic /d/ deletion by comparing corpora from two points in time. While the common practice has been to analyze data impressionistically in this investigation the dependent variable is defined as continuous by performing adapted acoustic measurements based on Carrasco, Hualde and Simonet's (2012) investigation. The findings reveal that while this phenomenon has been documented since the 17th century and it can be considered a vernacular variant in the speech community, in the most recent period, the sociolinguistic profile has changed. Variation is widespread across all age groups and socioeconomic classes.
While the probabilistic approach is certainly a norm in psycholinguistics, natural language proce... more While the probabilistic approach is certainly a norm in psycholinguistics, natural language processing and sociolinguistics, the categorical approach is still dominant in historical linguistics, describing only "the extremes" of the phenomenon and excluding the gradient continuum. Given that the period of language change may span over several hundreds of years, finding empirical evidence for language change with limited access to data presents a great challenge.
In this study, I offer a novel methodology to examine historical data, by combining methodologies from different fields, including probabilistic and corpus linguistic methods. First, this study focuses on infinitival transitive clauses, as it has been shown that the complexity of word order patterns and the structural ambiguity that obscure our analysis can be overcome through the examination of non-finite clauses. By applying corpus linguistic methods, the data are extracted from annotated corpora spanning over several centuries of Latin and Old French (PROEIL, Perseus, Opera Latina, MCVF and NCA). Second, the corpus is codified for various pragmatic, semantic and syntactic factors. The results of this study suggest that the origin of VO order seems to be in the reduction of the left periphery of clauses in Late Latin - a change that became evident only through analyzing reduced infinitival clauses.
This paper focuses on creating a historical parallel corpus of Old Occitan and English. The 13th ... more This paper focuses on creating a historical parallel corpus of Old Occitan and English. The 13th century Occitan narrative poem Roman de Flamenca holds a unique position in Provençal literature, and is a “universally acknowledged masterpiece of Old Occitan narrative” (Fleischmann, 1995). We show how historical investigations may benefit from such a parallel corpus
Aim: Psychosocial stress is known to affect substance use disorders. Little is known about how ps... more Aim: Psychosocial stress is known to affect substance use disorders. Little is known about how psychosocial stress relates to cognitive function and trajectories of opioid use disorder (OUD), specifically. Executive function (EF) is known to be impaired in substance abusers, and poor EF may adversely impact treatment adherence and treatment for OUD. The aim of this pilot study was to explore psychosocial factors associated with EF.
Methods: Community-based recruitment occurred using advertisements and referral from study participants. Eligibility included 18 years or older, opioid use in the past 90 days, and screening positive for DSM-5 OUD. Participants completed a questionnaire about substance use, social instability, food insecurity, drug use stigma, social support, perceived stress, depression, anxiety, and trauma. Executive function was assessed using a computerized version of the Iowa Gambling Test, a measure of risky decision-making, such that a higher proportion of advantageous selections (low risk, low reward) indicated better EF. Robust linear regression adjusting for age, race, education, and opioid severity was used to identify psychosocial factors associated with EF.
Results: The 46 participants were primarily male, white non-Hispanic, and >40 years; 89% used opioids at least weekly, 74% used within 48 hours of study participation, and polysubstance use was common. Higher anticipated drug use stigma was associated with worse EF (B= -3.05, p=0.01); higher emotional social support was associated with better EF (B=4.10, p=0.01). Higher food insecurity (B=0.43, p=0.01) and moderate cannabis use (B=6.04, p=0.03) were also associated with better EF.
Conclusion: Higher social support, lower stigma, and, paradoxically, more food insecurity and cannabis use were all strongly associated with better EF in individuals with OUD. Interventions that focus on social support and stigma could support improvements in EF and may improve OUD treatment retention (behavioral and pharmacological). Further studies are needed to clarify paradoxical findings.
American Evaluation Association Conference, Nov 14, 2019
The Complex Networks and Systems NSF Research Traineeship (CNS-NRT) is an interdisciplinary multi... more The Complex Networks and Systems NSF Research Traineeship (CNS-NRT) is an interdisciplinary multifaceted STEM program with a rigorous dual-PhD degree focus, emphasizing collaborative skills and workforce development. This complex institutional training model requires comprehensive evolutionary systems thinking and modeling, and novel data visualizations to communicate key results to diverse stakeholders. The Systems Evaluation Protocol (SEP) enables the inclusion of multiple perspectives, reflecting the complexity of program activities and outcomes. Through the SEP cyberinfrastructure platform, the Netway, the CNS NRT evaluation is visually modeled by logically connecting all program activities with short-, mid-, and long-term outcomes, building a comprehensive set of networked pathways that form the basis for an evaluation plan that integrates institutional data metrics and other measurement systems and enables comparison of CNS-NRT student outcomes with those of single PhD CNS and non-NRT dual PhD students. This paper describes the pathway model visualizations and the resulting comprehensive evaluation plan.
With advances in statistical computing and methods, there has been increasing interest in the use... more With advances in statistical computing and methods, there has been increasing interest in the use of advanced statistical models (e.g., mixed-effects modeling, conditional and random tree analyses) in sociolinguistic research. These advanced methods enable researchers to handle imbalanced data, measure individual and group variation, estimate significance, and rank variables according to their significance; their implementation, however, requires some programming skills or access to statistical tools that are not always freely available. To address these issues, we have developed a user-friendly interactive application Language Variation Suite (www.languagevariationsuite.com). Built as a Shiny app with various statistical and graphical R packages, LVS provides access to advanced statistical methods and visualization techniques to a broader audience, as its use does not require installation and programming skills.
In recent years, there has been growing interest in data visualization for text analysis. While... more In recent years, there has been growing interest in data visualization for text analysis. While text mining and visualization tools have evolved into mainstream research methods in many fields (e.g. social sciences, machine learning), their application to linguistic and literary studies still remains infrequent. First, many text processing tools require some programming skills, which take time to learn and are often challenging for digital humanities scholars. Secondly, while some visualization tools (e.g. Voyant, Weka and PaperMachine) provide graphical-user interfaces, social and humanities researchers seek more interactive and dynamic control of tools, which can serve as "holistic support for exploratory analysis" (Klein 2013). This workshop introduces two user-friendly applications, namely Language Variation Suite and Interactive Text Mining Suite, that allow researchers visually explore and statistically analyze language data. Written in R with Shiny app, these applications not only provide a web interactive interface, they also allow researchers implement state-of-the-art statistical methods, such as cluster analysis, topic modeling, inferential trees and mixed model logistic regressions.
Teaching Language and Teaching Literature in Virtual Environments, 2019
Recent advances in technology have made it possible to add immersive interactive dimensions to ne... more Recent advances in technology have made it possible to add immersive interactive dimensions to nearly any learning environment. This immersive technology provides students with active control and more authentic experiences; thus, helping them learn more effectively and increase their retention. In this view, these technologies seem to be an 'ideal' instrument for language instruction, as they combine visual, auditory, and kinesthetic learning styles. While their digital capabilities are almost limitless, their use in language learning remains limited due to technological and methodological challenges. This chapter provides methodological recommendations for the design and use of augmented and virtual technologies in language classroom settings. At the core of these recommendations is the collabora-tive research conducted at Indiana University which investigated the impact of mobile immersive technology for foreign language teaching and learning. Based on the findings, this chapter suggests several immersive tools and applications suited for the use in foreign language classroom (Aurasma, ThingLink, and Google Cardboard), which were evaluated by both students and instructors by means of self-assessment, technical feedback, and usage statistics.
Understanding the emergence, co-evolution, and convergence of science and technology (S&T) areas ... more Understanding the emergence, co-evolution, and convergence of science and technology (S&T) areas offers competitive intelligence for researchers, managers, policy makers, and others. The resulting data-driven decision support helps set proper research and development (R&D) priorities; develop future S&T investment strategies; monitor key authors, organizations, or countries; perform effective research program assessment; and implement cutting-edge education/training efforts. This paper presents new funding, publication, and scholarly network metrics and visualizations that were validated via expert surveys. The metrics and visualizations exemplify the emergence and convergence of three areas of strategic interest: artificial intelligence (AI), robotics, and internet of things (IoT) over the last 20 years (1998-2017). For 32,716 publications and 4,497 NSF awards, we identify their conceptual space (using the UCSD map of science), geospatial network, and co-evolution landscape. The findings demonstrate how the transition of knowledge (through cross-discipline publications and citations) and the emergence of new concepts (through term bursting) create a tangible potential for interdisciplinary research and new disciplines.
Uploads
Videos by Olga Scrivner
Brief introduction to AI, Unstructured data, Recommendation systems with an illustrative example of building Job Recommender system.
Psychosocial stress is a common factor contributing to the increase in opioid misuse. Stressors as
socio-environmental demands on individuals (e.g., trauma, adverse childhood events, unemployment,
stigma, food insecurity) have been shown to also impact cognitive processes. Particularly, the impact
has been noted on executive functioning (EF) linked to reasoning, problem-solving, decision-making.
Understanding the relationship between interpersonal trauma, opioid use disorder (OUD) and cognitive
impairments is essential for the rehabilitation trajectory and consequent improvement of social life.We assess 10 psychosocial stressors: social stability, food insecurity, substance use and severity, drug use stigma, social support, perceived stress, depression, anxiety, adverse childhood experiences, and trauma. Executive functions are
measured via two computer-assisted tests: Iowa Gambling Task, and Opioid Word Stroop Test.
Papers by Olga Scrivner
Brief introduction to AI, Unstructured data, Recommendation systems with an illustrative example of building Job Recommender system.
Psychosocial stress is a common factor contributing to the increase in opioid misuse. Stressors as
socio-environmental demands on individuals (e.g., trauma, adverse childhood events, unemployment,
stigma, food insecurity) have been shown to also impact cognitive processes. Particularly, the impact
has been noted on executive functioning (EF) linked to reasoning, problem-solving, decision-making.
Understanding the relationship between interpersonal trauma, opioid use disorder (OUD) and cognitive
impairments is essential for the rehabilitation trajectory and consequent improvement of social life.We assess 10 psychosocial stressors: social stability, food insecurity, substance use and severity, drug use stigma, social support, perceived stress, depression, anxiety, adverse childhood experiences, and trauma. Executive functions are
measured via two computer-assisted tests: Iowa Gambling Task, and Opioid Word Stroop Test.
Similar to a community-of-practice, the members of Scholarship of Teaching and Learning (SoTL) are driven by a shared interest and enthusiasm to improve their teaching and learning.
Representing time-varying SoTL events and relationships between SoTL members as a community network introduces challenges in data linking, data model, and network analysis. In particular, it is essential to design solutions to preserve the network topology, temporal information, member status transformation, and diverse relationships between nodes. In order to account for the SoTL network complexity, we design a heterogenous graph model and perform co-authorship and event network analyses to evaluate the effectiveness of the current SoTL strategies in attracting new members and supporting the sustainability of existing cohorts and provide data-driven decision support for SoTL programs in their development and priorities.
This workshop aims to provide a review of social media channels for classroom use. I will describe the cons and pros for most commonly used media and will discuss how to choose the right ones depending on the classroom goals.
In this study, I offer a novel methodology to examine historical data, by combining methodologies from different fields, including probabilistic and corpus linguistic methods. First, this study focuses on infinitival transitive clauses, as it has been shown that the complexity of word order patterns and the structural ambiguity that obscure our analysis can be overcome through the examination of non-finite clauses. By applying corpus linguistic methods, the data are extracted from annotated corpora spanning over several centuries of Latin and Old French (PROEIL, Perseus, Opera Latina, MCVF and NCA). Second, the corpus is codified for various pragmatic, semantic and syntactic factors. The results of this study suggest that the origin of VO order seems to be in the reduction of the left periphery of clauses in Late Latin - a change that became evident only through analyzing reduced infinitival clauses.
Methods: Community-based recruitment occurred using advertisements and referral from study participants. Eligibility included 18 years or older, opioid use in the past 90 days, and screening positive for DSM-5 OUD. Participants completed a questionnaire about substance use, social instability, food insecurity, drug use stigma, social support, perceived stress, depression, anxiety, and trauma. Executive function was assessed using a computerized version of the Iowa Gambling Test, a measure of risky decision-making, such that a higher proportion of advantageous selections (low risk, low reward) indicated better EF. Robust linear regression adjusting for age, race, education, and opioid severity was used to identify psychosocial factors associated with EF.
Results: The 46 participants were primarily male, white non-Hispanic, and >40 years; 89% used opioids at least weekly, 74% used within 48 hours of study participation, and polysubstance use was common. Higher anticipated drug use stigma was associated with worse EF (B= -3.05, p=0.01); higher emotional social support was associated with better EF (B=4.10, p=0.01). Higher food insecurity (B=0.43, p=0.01) and moderate cannabis use (B=6.04, p=0.03) were also associated with better EF.
Conclusion: Higher social support, lower stigma, and, paradoxically, more food insecurity and cannabis use were all strongly associated with better EF in individuals with OUD. Interventions that focus on social support and stigma could support improvements in EF and may improve OUD treatment retention (behavioral and pharmacological). Further studies are needed to clarify paradoxical findings.
Intro: https://languagevariationsuite.wordpress.com/2016/02/28/language-variation-suite/
Application: https://languagevariationsuite.shinyapps.io/Pages
To address these issues, we have developed a user-friendly interactive application Language Variation Suite (www.languagevariationsuite.com). Built as a Shiny app with various statistical and graphical R packages,
LVS provides access to advanced statistical methods and visualization techniques to a broader audience, as its use does not require installation and programming skills.
This workshop introduces two user-friendly applications, namely Language Variation Suite and Interactive Text Mining Suite, that allow researchers visually explore and statistically analyze language data. Written in R with Shiny app, these applications not only provide a web interactive interface, they also allow researchers implement state-of-the-art statistical methods, such as cluster analysis, topic modeling, inferential trees and mixed model logistic regressions.