At-risk Students Research Articles

ABSTRACT This study employs a multi-method approach, integrating quantitative and qualitative analyses to examine factors influencing student attrition at a community college in Trinidad and Tobago. Findings from the multilevel binary logistic regression model highlight a strong inverse relationship between GPA and attrition risk after adjusting for other variables in the model. This relationship varied significantly across the different academic programs. Additionally, being enrolled part-time, being older and being male were each associated with higher odds of withdrawal, controlling for the other covariates. Qualitative insights further contextualized these trends, with students’ frequently citing challenges related to program structure, institutional support, job conflicts, family obligations and financial constraints as key factors influencing their decision to withdraw. Based on these findings, early intervention programs, enhancedacademic advising, flexible course delivery, expanded financial aid, and targeted support forhigh-risk demographics are recommended for improving student retention. Leveraging a similardata-driven framework can help tertiary institutions proactively identify at-risk students and develop strategic, evidence-based retention initiatives that foster academic persistence and long-term student success.

To determine the prevalence and factors associated with depression, anxiety, and stress (DAS) in adolescents enrolled in secondary school. An institution-based cross-sectional study was conducted among 418 secondary school students who were selected using systematic sampling. The Depression, Anxiety, and Stress Scale (DASS-21) was used to collect data. Data were entered and analyzed by IBM SPSS version 30. Bivariable and multivariable analyses were performed to identify factors associated with depression, anxiety, and stress. In the final model, variables with a P-value <0.05 at 95% confidence intervals (CIs) were declared as statistically significant with DAS. A total of 418 study participants were included in the study, yielding a response rate of 99.3%. The overall prevalence of depression, anxiety, and stress in this study was found to be 123 (29.4%; 95% CI [25.0, 33.8%]), 199 (47.6%; 95% CI [42.8, 52.4%]), and 241 (57.7%; 95% CI [53.0, 62.4%]), respectively. The mean ± standard deviation of depression, anxiety, and stress scores were 7.74 ± 1.4, 6.27 ± 1.24, and 13.18 ± 1.76, respectively. In multivariable regression analyses, sex, living arrangement, substance use, connectedness with family, anxiety, and stress were found to be significantly associated with depression. Depression was found to be significantly associated with anxiety and living arrangement, level of education, connectedness with family, and family dispute, and anxiety was found to be significantly associated with stress. Overall, the results of the current study revealed that a remarkable proportion of students experienced depression, anxiety, and stress. Generally, socio-economic, academic, relationships with family, and substance use-related factors were identified as having an association with DAS. Strong relationships between depression, anxiety, and stress were discovered. Hence, we recommend structured counselling services for at-risk students for early detection and prevention of DAS.

Developing an assessment tool to identify mathematical misconceptions is important for early intervention and support for at-risk students. This exploratory sequential mixed methods study aimed to develop and validate a questionnaire for self-reflection on mathematical misconceptions among senior high school students using exploratory and confirmatory factor analyses as an application of structural equation modeling (SEM). This study involved 80 senior high school mathematics students across regions in the Philippines for the mathematical misconception test in the first phase. Of these, 20 purposively selected students who committed the most errors in the misconception test were interviewed to explore the underlying constructs of the students’ mathematical misconceptions. For the third and final phase, 310 selected students completed the developed self-reflected mathematical misconception scale. In this study, we identified four key factors of mathematical misconceptions: lack of procedural and conceptual knowledge, poor mathematical abstraction, internal barriers, and cognitive conflict. The developed scale, comprising 41 validated items, was tested valid and reliable tool for educators in assessing and addressing students’ mathematical misconceptions, allowing for designed instructional strategies and targeted interventions. Further research is recommended to explore the causes and remediation of mathematical misconceptions and track students’ progress in addressing them over time.

Hikikomori, a form of pathological social withdrawal or isolation, is increasingly recognized particularly among young people. This study examined the prevalence of subclinical hikikomori and Internet addiction and assessed their interrelationships from the perspective of network analysis among health-related university students. A cross-sectional study was conducted from September to December 2023 in China. Subclinical hikikomori and Internet addiction were assessed using the 1-month version of the 25-item Hikikomori Questionnaire (HQ-25M) and the Internet Addiction Test (IAT), respectively. Expected Influence (EI) and bridge EI were used as centrality indices to characterize the structure of the symptoms of both conditions. A total of 3,845 health-related university students were assessed. The prevalence of subclinical hikikomori and Internet addiction was 12.4% (95% confidence interval [CI] [11.4%, 13.4%]) and 26.4% (95% CI [25.0%, 27.8%]), respectively. The most central symptoms in this network model were "I avoid talking with other people" (HQ18), followed by "Request an extension for longer time" (IAT16) and "Academic efficiency declines" (IAT8). Additionally, "I spend most of my time alone" (HQ4), "Form new relationship" (IAT4) and "I do not like to be seen by others" (HQ11) were identified as bridge symptoms linking the communities of subclinical hikikomori and Internet addiction symptoms. Both subclinical hikikomori and Internet addiction were common among Chinese health-related university students. Interventions should target central and bridge symptoms of these conditions to reduce the comorbidity among at-risk university students.

High dropout rates continue to be one of the main barriers to the effectiveness of online learning. The objective of this study is to develop a predictive framework that identifies at-risk students early enough to enable timely intervention. The proposed approach relies on graph neural networks (GNNs) to capture how learners interact with digital resources over time. The learning environment is represented as a bipartite structure where students and course materials form nodes, and their connections are defined by frequency, type, and recency of interactions. The model was tested on a dataset of 3,000 students enrolled in 20 online courses over two academic semesters. A graph convolutional network (GCN) was implemented with embedding, layered convolution, dropout regularization, and a softmax output classifier. The results show that this framework outperforms commonly used models such as logistic regression, random forest, long short-term memory networks, and gradient boosting. It achieved strong predictive performance, with accuracy of 0.89, F1-score of 0.86, and area under the ROC curve of 0.91. In addition to improving predictive accuracy, the framework offers a dashboard that allows instructors to visualize learner engagement and detect borderline-risk profiles. These findings demonstrate that relational and temporal modeling with GNNs can provide a more reliable basis for early-warning systems, while supporting adaptive and learner-centered practices in digital education.

With the digital transformation in education, big data analytics is increasingly being used to understand, monitor, and improve students' academic performance. Analyzing student behavior, engagement levels, prior achievements, and study habits enables the creation of more effective and personalized learning environments. This study aimed to predict academic achievement from student data using machine learning (ML) algorithms and to identify the factors affecting achievement. Seven different algorithms were implemented for this purpose: SVM, LR, KNN, RF, NB, DT, and LDA. The RF, SVM, and LDA algorithms achieved the highest accuracy rate of 91%. The LDA model was determined to be the most successful model in terms of accuracy and balance performance. Analysis revealed that variables such as class participation, study time, and prior achievement level significantly impact student achievement. The findings demonstrate that self-management, self-regulation, and intrinsic motivation skills play a critical role in academic success. Consequently, machine learning-based models have strong potential for predicting student achievement and identifying at-risk students early. This study highlights the importance of data-driven decision-making processes in education and guides future research on AI-supported applications

This study investigates factors influencing student resignation and explores the role of academic departments and study programs in implementing early warning systems at Ciputra School of Business Makassar (CSB Makassar). Using a qualitative approach, in-depth interviews with students, faculty, and administrative staff were conducted to identify financial constraints, academic dissatisfaction, and social adaptation issues as key factors affecting retention. The research highlights the importance of proactive measures such as mentoring, academic advising, and personalized interventions to address these challenges. A proposed early warning system utilizing data analytics and real-time monitoring aims to identify at-risk students and deliver timely support. The findings contribute to understanding student retention dynamics and offer practical strategies for improving institutional practices. These insights are expected to enhance student engagement and satisfaction, providing a framework for other higher education institutions facing similar challenges.

The advancement in computing technology, online learning platforms, and pedagogical tools enable educators and learners to connect without temporal and geographical boundaries. The existing deep learning models to predict student performance are either simple recurrent neural networks or artificial neural networks employing demographic and hand-crafted features. This manuscript proposes a model, MultIFAR, that infuses multi-dimensional information representing different aspects of student behavior with an attention-driven deep learning model integrating bidirectional long short-term memory and convolutional networks to learn student representation efficiently. MultIFAR employs student demographic, assessment, and VLE-interaction to understand different aspects of student behavior from multifaceted data. MultIFAR includes bidirectional long short-term memory to process and capture patterns from demographic, assessment, and interaction information. The model applies a convolutional operation on the daily interaction information with the VLE. We also implement the attention mechanism to assign weight to significant features. The empirical evaluation over the Open University Learning Analytics (OULA) dataset establishes the efficacy of MultIFAR against the state-of-the-art approaches and baseline methods. Considering accuracy, MultIFAR reports results from 80.31% to 97.12% over the four different problems of student performance prediction. The ablation analysis reveals that diurnal interaction shows the highest, whereas demographic attributes show the least impact on MultIFAR accuracy. We also extend MultIFAR to predict at-risk and high-performing students early. We also evaluate the model over the balanced dataset and multiclass scenario.

The advent of big data and artificial intelligence (AI) technologies has catalyzed transformative changes across various sectors, including education. We propose two graph neural network-based models—MTGNN and DAGNN—for improving the prediction of student performance in educational settings. These models are evaluated using the Open University Learning Analytics Dataset (OULAD). MTGNN leverages multiple similarity metrics to build student relationship graphs that capture latent patterns, improving prediction outcomes. DAGNN enhances this approach by adding attention mechanisms and graph augmentation, leading to more accurate and robust predictions. Our experimental results show that both models outperform conventional baselines, especially in identifying at-risk students. This work demonstrates the potential of GNNs to transform educational analytics by modeling complex student data, enabling more effective, personalized interventions.

Predicting student academic outcomes is critical for enhancing personalized learning and enabling timely interventions for at-risk students. This study presents a comprehensive comparative evaluation of machine learning and deep learning models—specifically LightGBM, CatBoost, ANN, DNN, and WDNN—enhanced with ElasticNet and Lasso regularization to address challenges of high-dimensional educational data. Using the Math and xAPI datasets, thirteen AI models were evaluated through holdout and k-fold cross-validation across 100 iterations to assess predictive accuracy, generalizability, and interpretability. Results show that CNN with ElasticNet consistently achieves the highest accuracy (up to 93.67%), while ANN performs optimally with Lasso, demonstrating the effectiveness of regularization in improving model stability and reducing overfitting. The findings also highlight the practical utility of ensemble and deep learning models for early detection of at-risk students and support the development of explainable AI frameworks for educational analytics. By addressing prior research limitations, including narrow dataset scope and the absence of advanced hybrid models, this study advances scalable, interpretable, and reliable predictive systems, aiding institutions in data-driven educational decision-making.

The COVID-19 pandemic disrupted education systems in ways that continue to shape how learning, ability, and support are understood. While quantitative studies have repeatedly reported academic decline, particularly among at-risk students, there is still limited insight into how abilities were perceived in everyday teaching practice. This study draws on eight qualitative interviews with 12 teachers from inclusive and special schools in Vienna, Austria, conducted during and after school closures. Data were analysed using constructivist grounded theory ( Charmaz, 2014 ) combined with diffractive analysis ( Barad, 2007 ). Findings reveal five interrelated themes: the role of routine as a stabilising force, the emotional importance of physical closeness, the fragile stability of well-resourced students, opportunities for individual growth during lockdown, and enduring post-pandemic shifts in learning. Teachers’ narratives show that ability was constructed not as a fixed attribute but as contingent on context, expectations, and available support. While deficit-oriented views of students with special educational needs (SEN) persisted, some SEN students thrived under alternative conditions, challenging dominant learning loss discourses. The study highlights the need for flexible and inclusive educational frameworks that address structural ableism and can support diverse student needs, both in times of crisis and in everyday schooling.

The switch to online learning in higher education brought about by the Covid-19 pandemic has had lingering effects – most notably, continued higher levels of usage of learning management systems (LMS) such as Moodle for assessment and sharing of course materials. This has enhanced the potential for learning analytics even for courses that are delivered in a face-to-face mode. This is because the design of the course page on the LMS and how it is utilized for assessments over the semester necessarily affect the nature of student interactions with the LMS. There is already a sizeable literature that links student interactions with the LMS, selected student characteristics, and learning outcomes, highlighting that it is indeed possible to detect at-risk students using data sources such as course logs and click streams. However, there is less research on how early a student who is at risk of not completing or failing the course can be detected. This paper uses LMS logs, student characteristics, and learning outcomes of six cohorts of undergraduate students (over 500 students in total) taking a compulsory second-year module in a Sri Lankan university to detect the earliest point in the semester at which at-risk students can be identified. Due to the weekly modeling structure, the dataset expands to over 8,000 records, with each entry corresponding to a unique combination of student index number and week number. This paper employed a cumulative modeling approach, where several machine learning models including Random Forest, Decision Tree, Support Vector Machine (SVM), and K-Nearest Neighbors (KNN) are assessed for performance. Random Forest consistently outperformed other models, achieving an accuracy of 78.51% in Week 16. Notably, performance metrics stabilized above 70% by Week 8, suggesting it as the optimal point for early prediction. The analysis revealed that prior academic performance and consistency of LMS engagement were stronger predictors than total LMS clicks. These findings support the development of data-driven early warning systems tailored to the Sri Lankan higher education context, emphasizing the value of consistent behavioral monitoring and historical academic data for effective intervention strategies and it provides insights on how effectively utilizing an LMS can improve learning outcomes even for courses that are offered in face-to-face mode.

School connectedness encompasses protective elements that enhance students' academic and health outcomes. Its importance spans healthcare and education, yet clarity of the concept is needed as it relates to school nursing services. Using the Walker and Avant concept analysis method, this investigation aims to clarify the meaning of school connectedness, describe its uses, attributes, and characteristics in relation to school nursing services, and provide sample cases. The following definition emerged with relevance to school nursing services: School connectedness alludes to a student's perception of having a meaningful relationship with adults and peers within an inclusive and respectful school environment that supports the student's well-being, regardless of the student's health challenges. The goal is to cultivate a sense of fitting in for each child within their school community. Empowering school nurses to promote school connectedness can help identify at-risk students and foster a supportive, inclusive environment.

Using Machine Learning to Assess Factors Associated With North American Pharmacist Licensure Examination Performance.

Introduction: Hypertension is no longer a health issue limited to adults but increasingly affects adolescents. Its asymptomatic nature makes it difficult to detect early, while lifestyle factors such as poor diet, lack of physical activity, and stress significantly increase the risk. Early intervention through school-based programs is crucial to increase awareness and prevention. Objective: The purpose of this service was to provide promotive and preventive efforts against hypertension in adolescents at SMK Suryalaya, Tasikmalaya, by increasing their knowledge, awareness, and healthy lifestyle practices through health screening, education, and counseling Method: This public service was conducted by lecturers and students of Universitas Bakti Tunas Husada in collaboration with SMK Suryalaya. The program consisted of three stages: preparation, implementation, and evaluation. Screening included measurement of blood pressure, pulse, body mass index (BMI), and fitness tests. Health education was delivered through interactive lectures and discussions, while personal counseling was provided for students identified with health risks. Result: The activity involved 58 students (34 from grade X and 24 from grade XI). Screening results revealed 9.1% hypertension, 34.8% hypotension, and 13.6% obesity, indicating significant cardiovascular risks. Health education improved students’ knowledge with an average score increase from 60 (pre-test) to 82 (post-test), a gain of 36.7%, shifting most students from “moderate” to “high” knowledge categories. Personalized counseling for at-risk students enhanced their understanding and commitment to adopting healthier habits. Conclusion: This program demonstrated that integrating screening, education, and counseling effectively improved adolescent health awareness and promoted preventive behaviors. Early detection and health promotion in schools are essential strategies to reduce long-term cardiovascular risks. The program can serve as a replicable model for adolescent health promotion in other educational institutions.

Predictive validity of progress test scores and academic achievement on SMLE performance among Saudi medical graduates

In recent years, the integration of artificial intelligence (AI) in student management systems (SMS) has gained significant attention, particularly for monitoring academic performance and predicting at-risk students. Traditional approaches often lack the necessary adaptability and predictive accuracy across different learning environments. A hybrid AI-based model is proposed to enhance academic performance monitoring and intervention strategies by integrating decision trees (DT), random forests (RF), support vector machines (SVM), and artificial neural networks (ANN). The objective is to assess the effectiveness of the hybrid approach across multiple datasets, including UCI student performance, open university learning analytics dataset (OULAD), and national educational longitudinal study (NELS:88). The hybrid model was trained using a combination of preprocessing techniques, including missing data imputation, feature selection, and data normalization. The performance of the hybrid model was compared to individual base models using metrics such as accuracy, precision, recall, F1-score, and AUC-ROC. The hybrid model achieved outstanding results, with an accuracy of 98.8% on the UCI dataset, surpassing the performance of individual models. The hybrid model consistently outperformed the base models across all datasets, reducing error rates by over 5%. The proposed hybrid AI model provides a robust, scalable solution for academic performance monitoring and early intervention, demonstrating its potential for deployment in diverse educational contexts to support at-risk students proactively.

Conventional academic warning systems in higher education often rely on end-of-semester grades, which severely limits opportunities for timely intervention. To address this, our interdisciplinary study developed and validated a comprehensive socio-technical framework that integrates social-cognitive theory with learning analytics. The framework combines educational data mining with culturally responsive, personalized interventions tailored to a non-Western context. A two-phase mixed-methods design was employed: first, predictive models were built using Learning Management System (LMS) data from 2,856 students across 64 courses. Second, a quasi-experimental trial (n = 48) was conducted to evaluate intervention efficacy. Historical academic performance, attendance, and assignment submission patterns were the strongest predictors, achieving a Balanced Area Under the Curve (AUC) of 0.85. The intervention, specifically adapted to Confucian educational values, yielded remarkable results: 73% of at-risk students achieved passing grades, with a large effect size for academic improvement (Cohen’s d = 0.91). These findings empirically validate a complete prediction–intervention–evaluation cycle, demonstrating how algorithmic predictions can be effectively integrated with culturally informed human support networks. This study advances socio-technical systems theory in education by bridging computer science, psychology, and educational research. It offers an actionable model for designing ethical and effective early warning systems that balance technological innovation with human-centered pedagogical values.

The rapid advancement of technology has created new opportunities to enhance education, with machine learning (ML) emerging as a transformative tool. This study presents the development and evaluation of a comprehensive academic tracking system designed to monitor and categorize students based on performance metrics, while also providing functionality beyond simple grade reporting. Unlike traditional systems that serve primarily as repositories for academic scores, the proposed system offers integrated tools for tracking attendance, monitoring academic progress, managing assignments, and generating early alerts for at-risk students. Developed using Python for backend logic, React for frontend implementation, and MySQL for secure data management, the web-based platform was designed to improve real-time access and usability for both students and educators. The system incorporates a multifaceted methodology to analyze a wide range of student-related factors, including demographic data (e.g., age, gender, socioeconomic background), academic performance (e.g., grades, attendance), and behavioral indicators (e.g., participation and assignment submissions). The model classifies students into low, average, and high-performing groups using machine learning techniques, enabling more targeted interventions. When tested with real academic data from tertiary institutions in Nigeria, the proposed system demonstrated superior accuracy and efficiency in tracking and predicting student performance compared to existing solutions. These findings underscore the system’s potential to support data-driven decision-making in educational environments and to enhance learning outcomes through early identification and personalized support strategies.

Social media has become an integral component of student life, shaping communication, collaboration, and learning patterns. While it offers opportunities for academic engagement, excessive usage may lead to distractions and poor academic outcomes. This study investigates the impact of Social Media Usage (SMU) on Student Academic Performance (SAP), examining how demographic factors such as gender, department, and program type influence social media behaviour. Additionally, it explores the use of statistical and machine learning approaches to predict academic outcomes and develop AI-driven dashboards for identifying at-risk students. A structured questionnaire was administered to 150 students across selected colleges, yielding 135 responses. After validation for completeness and accuracy, 80 responses were considered suitable for analysis. A Simple Random Sampling (SRS) technique was employed to ensure unbiased representation. Collected data included demographic information, social media usage patterns, and academic performance indicators. Correlation analysis was applied to identify associations, ANOVA was used to examine demographic influences, and linear regression was employed to assess predictive effects. Machine learning techniques were integrated to enhance predictive accuracy and uncover latent patterns in the data. The findings indicate that SMU has a measurable impact on SAP, with both positive and negative influences depending on usage patterns. Gender, department, and program type were found to significantly moderate the relationship between social media engagement and academic outcomes. Furthermore, the AI-driven dashboards successfully identified students at risk of underperformance, providing a visual and actionable tool for educators to implement targeted interventions. This study offers important insights for students and educators seeking to balance social media engagement with academic achievement. By combining traditional statistical methods with machine learning and AI-driven visualization, the research demonstrates a practical approach to understanding and managing the complex relationship between social media behavior and academic performance. These results can inform strategies for promoting effective social media use and supporting at-risk students in higher education settings.