focalinterest.net

Tag: EdTech, Online Learning & Analytics

  • Text messaging programs support family literacy practices

    What the study found

    The review found that digital text messaging programs are under-used in preschool education, but they can help support families in young children's language and literacy learning. The authors also report that these programs may improve parent-child interactions, home literacy practices, and home-school transitions.

    Why the authors say this matters

    The authors conclude that digital messaging through short message services (SMS, text messages sent by phone) has not been fully used in early childhood education. They suggest these programs may be a highly effective way to support families and help with the transition between home and school.

    What the researchers tested

    This was a scoping review, using Arksey and O'Malley's 2005 approach, of current practices in text messaging programs that support families in preschool children's language and literacy learning. Searches of five databases identified 16 articles that met the review criteria.

    What worked and what didn't

    The review found text messaging programs were helpful in increasing parent-child interactions and supporting home literacy practices. It also found evidence that engaging parents and caregivers through text messaging can be an effective home-school transition strategy. The abstract does not report specific interventions that did not work.

    What to keep in mind

    This summary is based only on the abstract, so details about the included studies, their settings, and how the findings were assessed are limited. The abstract also notes a scarcity of peer-reviewed research in early childhood education, which means the evidence base described here is still small.

    • The review describes text messaging programs as an under-used tool in preschool education.
    • Sixteen articles met the review's inclusion criteria.
    • The abstract reports improvements in parent-child interactions and home literacy practices.
    • Text messaging was described as a potentially effective home-school transition strategy.
    • The abstract says there is a scarcity of peer-reviewed research in this area.

  • Deep learning classified GPA using family and psychological factors

    What the study found

    The study found that a deep learning framework combining family background and psychological evaluation indicators classified student GPA with the best performance when using TabTransformer plus a feature-gating mechanism. It also found negative associations between GPA and some psychological measures, including depression and anxiety.

    Why the authors say this matters

    The authors conclude that the framework may help identify academic risk and inform targeted academic assistance and psychological interventions. They suggest that using family background and psychological factors together can support this kind of classification in the post-pandemic era.

    What the researchers tested

    The researchers collected data from 1,692 undergraduates at a Chinese university. The dataset included family background factors, SCL-90 psychological evaluation scores, and GPA records, and they compared four deep learning models: TabTransformer, DCNv2, AutoInt, and MLP-ResNet, with a lightweight feature-gating mechanism added to improve feature selection.

    What worked and what didn't

    TabTransformer with the gating mechanism performed best, with an Accuracy of 0.798 and an AUC of 0.833. GPA was significantly negatively correlated with SCL-90 domains including depression and anxiety, and less favorable family background factors such as lower economic status and longer left-behind years were correlated with poorer psychological assessment outcomes.

    What to keep in mind

    The abstract describes data from one Chinese university, so the scope is limited to that sample. No other limitations are described in the available summary.

    • The best-performing model was TabTransformer with a gating mechanism.
    • Its reported performance was Accuracy 0.798 and AUC 0.833.
    • GPA was negatively correlated with SCL-90 domains including depression and anxiety.
    • Lower family economic status and longer left-behind years were linked to poorer psychological assessment outcomes.
    • The study used data from 1,692 undergraduates at a Chinese university.

  • AlphaLearn frames e-learning pathway design as multi-objective optimization

    What the study found

    The paper presents AlphaLearn, a conceptual framework for designing personalized e-learning pathways using evolutionary optimization. It argues that pathway design can be treated as a constrained multi-objective problem that balances learning effectiveness, efficiency, engagement, and fairness.

    Why the authors say this matters

    The authors conclude that equity should be part of adaptive pathway optimization, not something added afterward. The study suggests this is important because learner outcomes, failure rates, and dropout vary substantially across modules.

    What the researchers tested

    The researchers introduced a framework that combines knowledge graphs, learner modelling, and evolutionary algorithms to generate, evaluate, and refine candidate learning pathways. They also analyzed large-scale learning analytics data from the Open University Learning Analytics Dataset, or OULAD.

    What worked and what didn't

    The framework provides a structured description of the optimization cycle and pathway representation. The descriptive data analysis showed substantial variability in learner outcomes, failure rates, and dropout across modules, and the authors present fairness and bias mitigation as part of the framework; however, AlphaLearn is described as conceptual and methodological rather than a validated system.

    What to keep in mind

    The abstract does not describe empirical validation of AlphaLearn, so its performance is not established here. The summary also presents descriptive findings from OULAD, but does not report experimental comparisons or outcome metrics for the framework itself.

    • AlphaLearn is presented as a conceptual evolutionary framework for personalized e-learning pathways.
    • The framework treats pathway design as a constrained multi-objective optimization problem.
    • Fairness and bias mitigation are described as integral to the optimization process.
    • A descriptive analysis of OULAD showed substantial variability in learner outcomes, failure rates, and dropout across modules.
    • The paper does not present AlphaLearn as a validated system.

  • Survey finds 172 open datasets in learning analytics papers

    What the study found

    The study found 172 unique open datasets associated with 204 publications in learning analytics, educational data mining, and AI in education. The authors present this as the most comprehensive collection and analysis of open educational datasets to date.

    Why the authors say this matters

    The authors say open datasets support reproducibility, collaboration, and trust in research findings, and can also increase authors' visibility, credibility, and citation potential. They conclude that their findings and checklist may support wider adoption of open data practices in learning analytics communities and beyond.

    What the researchers tested

    The researchers conducted a systematic survey of publicly available datasets published alongside research papers. They manually examined 1,125 papers from three flagship conferences, LAK, EDM, and AIED, covering the years 2020 to 2024.

    What worked and what didn't

    The survey identified 172 datasets, and 143 of them were not captured in any prior survey of open data in learning analytics. The authors also categorized the datasets and analyzed their context, analytical methods, use, and other properties, and they summarized current gaps in the field.

    What to keep in mind

    The available abstract does not describe detailed study limitations beyond the conference and year range examined. The survey covers only publicly available datasets tied to papers in the three named flagship conferences from 2020 to 2024.

    • 172 unique datasets were identified across 204 publications.
    • The survey covered 1,125 papers from LAK, EDM, and AIED between 2020 and 2024.
    • 143 of the identified datasets were not included in prior surveys.
    • The authors provided an annotated inventory of datasets and related publications.
    • The paper includes an 8-item PRACTICE checklist for publishing data.

  • Parental AI investment is linked to greater AI-mediated English learning

    What the study found

    The study found that parental AI investment behaviours were directly associated with students' engagement in AI-mediated informal digital learning of English (AI-IDLE). It also found indirect effects through children's perceived AI value and effort expectancy for AI, including a chain mediation pathway.

    Why the authors say this matters

    The authors conclude that the study extends Situational Expectancy-Value Theory to informal, self-regulated, technology-enhanced language learning contexts. They also say it highlights the role parents play in out-of-class learning and offers practical insights for guiding children to use AI tools for English proficiency development.

    What the researchers tested

    The researchers used a questionnaire survey with 2,346 primary and secondary school students in China. They analyzed the data with structural equation modelling to test direct effects and mediating effects between parental AI investment behaviours, perceived AI value, effort expectancy for AI, and AI-IDLE engagement.

    What worked and what didn't

    Parental AI investment behaviours were reported to promote AI-IDLE engagement directly. The study also found significant indirect effects through perceived AI value alone, effort expectancy for AI alone, and a chain mediation path involving both variables.

    What to keep in mind

    The abstract does not describe specific limitations. The study reports associations from a questionnaire survey of students in China, so the summary available here does not provide information about causation or broader generalizability.

    • Parental AI investment behaviours were directly linked to higher AI-IDLE engagement.
    • Perceived AI value mediated the relationship between parental AI investment and AI-IDLE engagement.
    • Effort expectancy for AI also mediated the relationship.
    • A chain mediation pathway through perceived AI value and effort expectancy for AI was found.
    • The study surveyed 2,346 primary and secondary school students in China.

  • Students valued an AI learning assistant but had ethical concerns

    What the study found

    Students generally valued the Educational AI Hub, an AI-powered learning framework, for being accessible and comfortable to use. Many saw it as helpful for homework and understanding concepts, but they also expressed ethical uncertainty, especially about institutional policy and academic integrity.

    Why the authors say this matters

    The study suggests that usability, ethical transparency, and faculty guidance are important for meaningful AI engagement in higher education. The authors conclude that students viewed AI as a supplement rather than a replacement for human instruction.

    What the researchers tested

    The researchers evaluated the Educational AI Hub in undergraduate civil and environmental engineering courses at a large U.S. public university. They used a mixed-methods design with pre- and post-surveys, system usage logs, and qualitative analysis of students' AI interactions.

    What worked and what didn't

    The AI assistant was most helpful for completing homework and understanding concepts. Nearly half of students said it was easier to use than asking instructors or teaching assistants for help, but views on its instructional quality were mixed and ethical uncertainty was a barrier to fuller engagement.

    What to keep in mind

    The study involved 71 students across two courses at one large public university, so the findings are limited to that setting. The abstract also does not describe detailed limitations beyond the scope of the sample and context.

    • Students valued the AI assistant for accessibility and comfort.
    • Nearly half said it was easier to use than asking instructors or teaching assistants for help.
    • The tool was most helpful for homework and concept understanding.
    • Ethical uncertainty about policy and academic integrity limited engagement.
    • Students treated AI as a supplement, not a replacement, for human instruction.

  • VR and smart home use were linked to greater family intimacy

    What the study found

    The study found that virtual reality (VR) and smart home technologies were each indirectly associated with higher family emotional intimacy through different pathways. VR was linked through increased parenting self-efficacy, meaning parents' confidence in their parenting, while smart home technologies were linked through reduced parenting burden.

    Why the authors say this matters

    The authors conclude that the Technology-Based Family Resource Model supports VR and smart home technologies as complementary psychosocial resources for families. They also suggest the findings support gender-neutral, technology-integrated family interventions and provide a theoretical basis for digital equity policies that support early parental well-being.

    What the researchers tested

    The researchers tested the Technology-Based Family Resource Model, which proposes two pathways: a resource acquisition pathway through VR and parenting self-efficacy, and a resource conservation pathway through smart homes and reduced parenting burden. They studied 169 parent couples with children aged five years and under, using quantitative analyses including confirmatory factor analysis, structural equation modeling, and multi-group structural equation modeling, plus a qualitative component with 20 participants.

    What worked and what didn't

    The structural model fit the data well, with CFI = .94 and RMSEA = .04. VR utilization significantly increased family intimacy indirectly through parenting self-efficacy (β = .17, p < .001), and smart home utilization significantly increased intimacy indirectly through reduced parenting burden (β = .11, p < .001). Multi-group analysis showed full structural invariance across fathers and mothers (ΔCFI = .002, p > .05), indicating the model worked similarly for both groups.

    What to keep in mind

    The abstract does not describe specific limitations beyond the study scope. The sample included parent couples of children aged five years and under, so the findings are limited to that group.

    • Virtual reality was indirectly linked to greater family emotional intimacy through higher parenting self-efficacy.
    • Smart home technologies were indirectly linked to greater family emotional intimacy through lower parenting burden.
    • The statistical model fit well and showed the same structure for fathers and mothers.
    • The study included 169 parent couples with children aged five years and under, plus a qualitative subsample of 20 participants.

  • Parental AI literacy is central in early childhood education

    What the study found

    The review argues that parents and families play a crucial role in how artificial intelligence (AI) is used in early childhood education. It presents a conceptual framework for parental AI literacy with three parts: AI knowledge, AI skills, and AI attitudes.

    Why the authors say this matters

    The authors state that understanding parental AI literacy is essential for shaping children's learning experiences and developmental outcomes. They also suggest that family-centered approaches may help use AI's advantages while reducing potential risks.

    What the researchers tested

    This paper is a review of existing literature on AI in early childhood education (ECE), meaning education for young children before primary school. The review also proposes a new conceptual model for parental AI literacy in the family context.

    What worked and what didn't

    The paper classifies AI applications in ECE into six categories: interactive AI, generative AI, AI prediction, AI literacy, AI-driven personalized learning, and affective AI. It says these developments offer advantages, but they also raise challenges including digital disparities, data privacy concerns, and ethical dilemmas.

    What to keep in mind

    The abstract does not report original empirical data or test the proposed framework. It also does not give detailed study limitations beyond noting that the framework is intended as a foundation for future research.

    • The review emphasizes parents and families as important in children's AI-related learning in early childhood education.
    • It proposes a three-part framework for parental AI literacy: knowledge, skills, and attitudes.
    • The abstract says parental mediation strategies can vary with socioeconomic status, cultural background, age, and education.
    • AI in early childhood education is grouped into six categories, including generative AI and AI-driven personalized learning.
    • The paper notes possible challenges such as digital disparities, privacy concerns, and ethical dilemmas.

  • AI tools improve student engagement when paired with teaching methods

    What the study found

    The review found that AI tools in higher education are most effective for student engagement when they are used with interactive teaching methods. The authors also propose the PMAISE model, which stands for Pedagogical Mediation of AI for Student Engagement, to describe how AI, pedagogy, and engagement are connected.

    Why the authors say this matters

    The authors conclude that AI in higher education should be integrated in a context-sensitive, evidence-based, and pedagogically meaningful way. They suggest that thoughtful pedagogical mediation is crucial for maximizing AI’s educational benefits.

    What the researchers tested

    The researchers conducted a systematic review of 73 peer-reviewed articles published between 2015 and early 2025. They searched Scopus and Web of Science, followed PRISMA guidelines, and coded studies using a framework covering AI types, engagement outcomes, and instructional strategies.

    What worked and what didn't

    AI tools such as chatbots, adaptive systems, and predictive analytics were reported to enhance engagement most effectively when paired with flipped classrooms, project-based learning, and scaffolded feedback loops. The review also notes that teaching methods can amplify or inhibit the effects of AI tools, and it highlights concerns related to ethics, data privacy, and structural barriers to equitable AI adoption.

    What to keep in mind

    This is a review of previously published studies, not a single new experiment. The abstract does not provide detailed limitations beyond noting concerns about ethics, data privacy, and barriers to equitable adoption.

    • The review analyzed 73 peer-reviewed articles from 2015 to early 2025.
    • AI tools were most effective for engagement when used with interactive pedagogies.
    • The paper introduces the PMAISE model to map AI, pedagogy, and engagement.
    • Chatbots, adaptive systems, and predictive analytics were among the AI tools discussed.
    • The abstract highlights ethics, data privacy, and equity barriers as concerns.