This study investigated first‐year undergraduate statistics students’ statistical problem‐solving skills on the probability of the union of two events, conditional probability, binomial probability distribution, probabilities for x‐limits using the z‐distribution, x‐limit associated with a given probability for a normal distribution, estimating the y‐value using a regression equation, and hypothesis testing for a single population mean when a population standard deviation is unknown. The study was a descriptive case study and employed a mixed‐method research approach. Data were collected through content analysis of a statistics course examination script of 120 first‐year undergraduate students of statistics in an open distance‐learning university in South Africa. Polya's Model of Problem Solving was used as the framework of analysis. The study revealed that the students, in general, had poor statistical problem‐solving skills.
17 October 2026 in South African Journal of Higher Education
Artificial intelligence is reshaping assessment practices in higher education, yet its ethical governance in Global South open distance e-learning (ODeL) contexts is complicated by persistent inequalities in digital access, skills, and institutional capacity. This article examines how AI-related academic integrity challenges emerged in a large undergraduate data handling and probability module within a South African teacher education programme offered through an ODeL institution. Framed by Sociotechnical Systems Theory, the study analyses assessment submissions from a cohort of 3,088 students and a purposive sample of 229 submissions, triangulated with written reflections from experienced module markers. The findings identify recurring integrity challenges that cannot be reduced to individual misconduct, including AI-shaped graphical artifacts and procedural calculations, handwritten and scanned submissions that bypass similarity checking, and non-acceptance of End-User License Agreements that weaken accountability. Marker evidence shows how these practices are amplified by weak digital literacies, device constraints, and equity-driven accommodations that unintentionally create loopholes in integrity systems. The article argues that ethical AI governance in resource-constrained ODeL environments requires alignment among policy, technology, and pedagogy, including inclusive integrity pathways, strengthened support for digital and AI literacy, and assessment designs that foreground reasoning, conceptual understanding, and pedagogical interpretation.
Digital assessment has become central to Open Distance e-Learning (ODeL), yet its implementation often assumes equitable access to digital resources, potentially disadvantaging students studying in diverse geographical and socio-economic contexts. This study examined how students’ living contexts influence their experiences of digital assessment in undergraduate mathematics education at a South African distance education institution. An explanatory sequential mixed-methods design was employed. Survey data were collected from 253 undergraduate students, with 251 participants included in the inferential analyses after excluding two respondents who selected “Other” as their living area category. Quantitative data were analysed using chi-square tests, binary logistic regression, and ordinal logistic regression, while qualitative responses were analysed thematically. The findings revealed that living context was significantly associated with reliable internet access, and internet reliability significantly associated with students’ experiences of typing and submitting digital assessments as well as their perceptions of assessment inclusivity. Qualitative findings further highlighted challenges relating to connectivity, device limitations, mathematical representation, and the need for more flexible assessment practices. The study advances Contextualized Digital Assessment as a conceptual framework for designing equitable digital assessment that recognises students’ diverse technological, geographical, disciplinary, and socio-economic realities. The findings provide practical guidance for higher education institutions seeking to design more inclusive and context-sensitive digital assessment within ODeL environments.
This chapter examines how assessment can be reimagined in HyFlex environments in response to the growing influence of artificial intelligence (AI). It argues that AI should not be positioned primarily as a tool for automation or surveillance, but as a carefully governed support for authentic, equitable, and pedagogically meaningful assessment. Drawing on TPACK and PESTEL as complementary analytical lenses, the chapter explores the challenges of multimodal assessment, including equivalency, workload, access, governance, and institutional readiness. It further proposes that authentic AI-supported assessment in HyFlex depends on the alignment of human academic judgement, equity-conscious design, and institutional capacity. In doing so, the chapter offers a practice-informed conceptual contribution with illustrative South African insights for institutions seeking to move from automation-centred assessment toward more transparent, responsive, and sustainable assessment practices. .