The GPhase workflow proceeds through three steps. First, the unitig graph is partitioned into chromosome-scale subgraphs by cutting specific edges, with subgraphs grouped into clusters on the basis of k-mer similarity and then assigned to chromosome bins using global 3C interaction signals. Second, within each chromosome bin, unitigs are separated into haplotypes. Collapsed unitigs—regions where homologous sequences have been merged into a single node—are used to guide phasing: the algorithm first clusters the non-collapsed unitigs directly connected to each collapsed unitig, then validates the clustering by examining the degree of sequence homology within each cluster, and finally assigns the collapsed unitigs themselves according to the strength of 3C signals and sequence homology to the non-collapsed unitigs in each haplotype. Third, within each haplotype, unitigs are linearized using local graph topology, and disjointed sequences are ordered and oriented with global 3C linkage information to yield final chromosome-scale scaffolds.

Benchmarking on simulated polyploid genomes with varying degrees of sequence similarity, GPhase consistently outperformed existing tools, reducing the sequence missing rate by up to 22.5-fold and the sequence error rate by up to 12.5-fold. On real genomes—including potato, rose, and kiwifruit—GPhase produced chromosome-level assemblies with fewer misassemblies and higher contiguity than competing approaches. The algorithm also recovered haplotype-resolved sequences that existing pipelines missed entirely, demonstrating its capacity to resolve the complex homologous relationships that characterize polyploid genomes.

The study, published in Nature Plants, provides both the algorithm and a practical implementation, offering a broadly applicable solution for polyploid genome assembly. By unifying assembly graphs and 3C signals in a single framework, GPhase addresses a long-standing gap in genomics and opens the door to more accurate and complete assemblies of the many crop species whose genomes are polyploid.

The scale of the polyploid assembly problem is easy to underestimate. Roughly 70% of flowering plants carry evidence of polyploid ancestry in their evolutionary history, yet polyploid species make up only about 18% of the high-quality genome assemblies currently available. That gap is not a matter of neglect. It reflects a technical bottleneck: when three, four, or more copies of every chromosome coexist in a single nucleus and share stretches of highly similar sequence, the computational machinery that cleanly separates maternal from paternal chromosomes in diploids struggles to tell homologs apart. Larger genome sizes and higher proportions of repetitive elements compound the difficulty, which is why many economically and scientifically important polyploid crops still lack chromosome-scale, haplotype-resolved reference genomes.

The two-step structure of modern polyploid assembly pipelines helps explain where things go wrong. The first step, in which long reads are assembled into unitigs, is essentially a solved problem: general-purpose assemblers such as hifiasm, Verkko, Flye, and NextDenovo handle it well, and they routinely output assembly graphs as a byproduct. The second step—assigning unitigs to haplotypes and arranging them into chromosome-scale scaffolds—is where polyploid projects stall. This step depends on independent linkage information, and the dominant source has been chromatin conformation capture data, prized for its favorable balance of cost, scalability, and effectiveness. Hi-C remains the most widely used flavor, with newer approaches such as Pore-C and Omni-C gaining adoption because their longer reads or freedom from restriction enzyme cut sites offer theoretical advantages.

But 3C data share a fundamental weakness regardless of flavor: they provide only coarse-grained linkage. Two failure modes follow. Short unitigs that happen to lack restriction cut sites, or that fall in regions of uneven sequencing coverage, produce sparse or absent interaction signals and are often discarded outright, leaving chromosomes incomplete. Meanwhile, homologous regions with high sequence similarity are frequently collapsed into consensus sequences during unitig assembly, and these collapsed unitigs emit strong interaction signals toward multiple haplotypes at once, corrupting the clustering step and producing uneven, mis-phased assemblies. Existing polyploid scaffolding tools, ALLHiC and HapHiC, were designed around Hi-C data and have powered most recent polyploid assemblies, but they cannot escape these data-inherent limitations.

The central insight behind GPhase is that the assembly graph, which is already generated by mainstream unitig assemblers, contains exactly the complementary information that 3C data lack. Because graph edges record overlap relationships between long reads, they carry fine-grained connectivity at the sequence level. A short unitig with no usable Hi-C signal can still be placed correctly through its graph connections to neighbors. A collapsed unitig betrays itself in the graph by forming links to multiple haplotypes simultaneously, allowing the algorithm to detect and handle it rather than letting it mislead the phasing process. The graph topology even hints at the relative ordering of unitigs along a chromosome, and it enables something no existing polyploid scaffolding tool offers: the direct assembly of unitigs into longer contigs, which boosts the contiguity of gap-free sequence.

The benchmarking design deserves attention because it separates what an algorithm claims from what it can prove. The evaluation set included two semi-simulated polyploid genomes with high-quality ground-truth assemblies, constructed by merging four and three homologous haploid rice genomes respectively—a clever trick that provides perfect known answers against which phasing accuracy can be measured. Four purely in-silico-simulated autopolyploid genomes with varying degrees of sequence similarity probed how performance degrades as homologs become harder to distinguish. Five real polyploid genomes, including potato, wax apple, and rose, tested performance under messy real-world conditions, and an additional four polyploid genomes served as an independent validation set. Across Hi-C, Pore-C, and Omni-C data types, GPhase reduced the BUSCO-based sequence missing rate by up to tens of times, cut the k-mer-based sequence error rate by up to an order of magnitude, and improved the N50 contiguity of gap-free sequences by several-fold relative to existing tools.

These metrics map onto distinct biological consequences. A reduced missing rate means fewer genes and regulatory elements are absent from the final assembly, which matters enormously for gene annotation, comparative genomics, and any analysis that depends on knowing what is present. A lower error rate means the assembled sequence more faithfully represents the actual chromosomes, reducing false variants and spurious structural differences that would otherwise propagate into downstream studies. Improved contiguity of gap-free sequence means fewer unresolvable breaks, which preserves synteny information and makes it easier to trace how genes and genome regions have been rearranged, duplicated, and retained or lost across evolutionary history.

The distinction between autopolyploids and allopolyploids, though not the focus of the algorithm itself, underlies why sequence similarity among homologs is such a decisive variable. Autopolyploids arise when genome copies derive from the same species, so their homologous chromosomes can be nearly identical at the sequence level, making discrimination hardest. Allopolyploids combine subgenomes from different ancestral species, giving each subgenome a distinct sequence identity that partially eases phasing. By testing on simulated genomes with graded levels of similarity, the benchmark captured this spectrum directly, and the consistent advantage of GPhase across that spectrum suggests the assembly-graph signal remains informative even when 3C signals alone would be hopelessly confounded.

It is also notable that diploid-aware tools that already exploit assembly graphs, such as hifiasm, do not simply extend to polyploid genomes. This indicates that the polyploid problem is not merely a diploid problem with more copies; the combinatorics of assigning unitigs among three or more haplotypes, the prevalence of collapsed regions, and the ambiguity of multi-haplotype 3C signals create qualitatively new challenges that require purpose-built algorithmic treatment. GPhase’s preprocessing normalizes and unifies 3C input, its partitioning step cuts the unitig graph into chromosome-scale subgraphs, its phasing step uses collapsed unitigs as anchors to guide local clustering with validation by homology degree, and its scaffolding step linearizes unitigs within each haplotype using local graph topology before applying global 3C linkage to order and orient disjointed sequences.

For the plant genomics community, the practical implications are broad. Many staple and specialty crops—potato among the most prominent—are polyploid or have polyploid ancestors, and their improvement depends on understanding trait architecture, subgenome dominance, and heterosis, all of which require knowing which alleles sit on which homolog. Haplotype-resolved, chromosome-scale assemblies are the substrate for that knowledge. By making such assemblies more accurate, more complete, and more contiguous, and by doing so with data types that laboratories already generate, an approach like this lowers the barrier for species that have waited longest for reference-quality genomes. The open availability of both the algorithm and its implementation further suggests that the gap between the polyploid share of flowering plants and the polyploid share of high-quality assemblies may finally begin to close.

Subject of Research: Genomics

Article Title: Assembly-graph-assisted haplotype-resolved scaffolding of polyploid genomes

Article References: Du, W., Li, Y., Zeng, X., Yang, J., Wang, X., Zhang, X., Shang, L., Fu, X., Zhang, X., Wei, X.-Q., Cheng, T., Li, D., Wu, Z., & Pan, W. (2026). Assembly-graph-assisted haplotype-resolved scaffolding of polyploid genomes. Nature Plants. https://doi.org/10.1038/s41477-026-02422-5

Image Credits: AI Generated

DOI: 10.1038/s41477-026-02422-5

Keywords: GPhase, polyploid genomes, assembly graphs, 3C signals, haplotype-resolved assembly, chromosome-level scaffolding, Nature Plants, Assembly-graph-assisted, haplotype-resolved, scaffolding, polyploid, genomes

The research team, led by John Ackon with colleagues Francis Ohene Boateng and Benjamin Adu Obeng, worked with 120 Form Two students at Asanteman Senior High School, a population of roughly 800 students at the stage where algebra becomes a gateway to science, engineering, technology and economics. Two intact classes of 60 students each were compared. One class received the experimental treatment; the other continued with the lecture method that still predominates in many Ghanaian senior high schools. Both groups covered the same algebra topics, including algebraic expressions, linear equations and contextualised problem-solving, and both received equal instructional time from the same teacher, a design choice intended to isolate the instructional strategy as the only meaningful difference between conditions.

The mechanics of the intervention are worth examining because they illustrate what differentiated instruction actually looks like when it moves from theory to practice. Before lessons began, students’ readiness levels were diagnosed using a pretest. Those with weaker foundations received additional teacher guidance, worked examples, step-by-step explanations and guided practice, while stronger students were given more challenging tasks designed to stretch independent reasoning. Each lesson followed a structured scaffolding arc: explicit explanation and modelling first, then guided practice with questioning, peer collaboration and differentiated tasks, then a deliberate fading of support. As students demonstrated growing competence and confidence, the teacher reduced prompts, increased independent practice and allowed learners to tackle progressively harder problems with less assistance.

This fading is the conceptual heart of scaffolding. Drawing on Vygotskian ideas, the approach targets what psychologists call the Zone of Proximal Development, the space between what a learner can do alone and what they can do with help. Support is temporary by design; the goal is not dependence on the teacher but a progressive transfer of responsibility to the student. The researchers also drew on Multiple Intelligences Theory, which holds that learners possess varied intellectual strengths and therefore benefit from varied instructional pathways, a framework that remains debated among psychologists but has proven practically useful for justifying differentiated classrooms.

The outcome measure was a researcher-developed Mathematics Achievement Test of ten application-based algebra questions, designed to capture both mathematical understanding, meaning the interpretation of expressions and identification of relationships, and problem-solving performance, meaning the application of strategies and procedures to algebraic situations. The instrument was vetted by experts in mathematics education and measurement, pilot-tested with 115 comparable students, and showed acceptable psychometric credentials: an average item difficulty of 0.59, an average discrimination index of 0.43, and internal consistency of 0.79 on the Kuder-Richardson Formula 20. Student perceptions were measured with a ten-item questionnaire on a five-point scale, which achieved a Cronbach’s alpha of 0.764.

The results were striking. Before the intervention, the two groups were statistically indistinguishable, with mean pretest scores of 61 for the experimental group and 60 for the control group. After three weeks, the lecture group had improved modestly from 60 to 65, a five-point gain. The scaffolding-based differentiated instruction group climbed from 61 to 75, a fourteen-point gain. Because the researchers used Analysis of Covariance with pretest scores as a covariate, they could statistically adjust for any initial differences, and the adjusted comparison was emphatic: instructional approach had a significant effect on posttest performance, F(1,117) = 278.511, p < .001, with a partial eta squared of 0.704, an unusually large effect size for a classroom intervention. The overall model explained 74.1 percent of the variance in posttest scores. Students taught by lecture scored, on average, 8.90 points lower than their scaffolded peers after adjustment, with a 95 percent confidence interval of 6.35 to 11.45 points.

Perhaps more telling than the raw means was the movement of students across proficiency categories. The researchers classified students as Approaching Proficiency, Proficient or Highly Proficient based on percentage scores. Before the intervention, most students in both groups sat in the lowest category, with 24 students in the experimental group and 23 in the control group below the 60 percent threshold. Afterward, the lecture group’s lowest band shrank only slightly, from 23 to 18 students. In the experimental group, it collapsed from 24 to just 5. Meanwhile, the number of Highly Proficient students in the scaffolded class more than doubled, from 13 to 29, while the control group managed only 13 to 15. This pattern suggests the approach did not merely lift the middle; it pulled struggling students across a meaningful threshold and pushed capable students toward genuine mastery, exactly the dual effect differentiated instruction is designed to achieve.

Students noticed the difference. The perception questionnaire, administered only to the experimental group, yielded an overall mean of 3.423 with a standard deviation of 0.588, indicating moderately positive views. The highest-rated items, each scoring 3.54, were that step-by-step support helped students understand better and that they preferred the method to the usual lecture approach. Students also rated hints and guidance for difficult questions, peer collaboration, and growing independence in problem-solving favourably. The lowest-rated item, at 3.28, was the statement that support was gradually reduced as students improved, still above the scale midpoint but hinting that the fading of help, the most theoretically crucial element, may be the hardest part of scaffolding for students to perceive or appreciate.

The authors are appropriately careful about what these numbers can and cannot prove. Because intact classes rather than randomly assigned individuals were used, instructional condition was completely confounded with class membership, meaning pre-existing classroom dynamics could partly explain the gap, even though ANCOVA controlled for baseline achievement. Both conditions were taught by the same teacher, which preserved consistency but raises the possibility of expectancy effects. The study took place in a single school with convenience sampling, the intervention lasted only three weeks, and perception data came solely from the treated group, so claims about motivation, confidence or engagement remain interpretive rather than demonstrated. The researchers also note they did not directly measure engagement or depth of conceptual understanding, so the mechanisms behind the achievement gains, while plausible and theory-consistent, are not confirmed.

Even with those caveats, the study matters. It provides experimental evidence from a context where teacher-centered instruction persists despite national policies, including Ghana’s Education Strategic Plan and the National Pre-Tertiary Education Curriculum Framework, that explicitly call for learner-centered, competency-based teaching. It moves beyond the simple pre-post comparisons that dominate the differentiated instruction literature by tracking proficiency-level transitions and student perceptions. And it suggests a concrete, replicable recipe for algebra teachers: diagnose readiness first, match tasks and support to each level, use worked examples, questioning, peer collaboration and continuous formative feedback, then deliberately withdraw help as competence grows. The authors call for longer interventions, follow-up assessments to test whether gains persist, multiple teachers, and multi-school studies to establish generalizability. But the core message is already legible: when algebra instruction meets students where they are and scaffolds them toward independence, achievement follows, and the students furthest behind may benefit most.

Subject of Research: The effect of scaffolding-based differentiated instruction on algebra achievement, mathematical understanding and problem-solving performance among senior high school students

Article Title: Scaffolding-based differentiated instruction and students’ algebra achievement: evidence from mathematical understanding and problem-solving performance

Article References: Ackon, J., Boateng, F. O., & Obeng, B. A. (2026). Scaffolding-based differentiated instruction and students’ algebra achievement: evidence from mathematical understanding and problem-solving performance. Frontiers in Education, 11, Article 1954830. https://doi.org/10.3389/feduc.2026.1954830

Image Credits: AI Generated

DOI: 10.3389/feduc.2026.1954830

Keywords: algebra achievement, differentiated instruction, scaffolding, mathematics education, problem-solving performance, Zone of Proximal Development, quasi-experimental design, ANCOVA, Ghana, senior high school, learner-centered pedagogy, student perceptions

Shared reading has long been recognized as a powerful engine for early language development. Unlike everyday conversation, which tends to recycle familiar vocabulary, books introduce children to novel words that rarely surface in routine exchanges. The dialogic nature of shared reading—parents and children co-constructing narratives from illustrations—gives children the chance to stretch beyond their current vocabulary through interactions with more knowledgeable partners. Yet despite a growing literature on shared reading, most of it has focused on monolingual families or relied on surveys and controlled experiments. How bilingual parents and children actually use their two languages during natural reading interactions at home has remained largely uncharted territory.

The research team, based at the University of Oslo, grounded their work in a social-interactionist framework, which holds that children’s language learning emerges through social interaction. They focused on two distinct ways parents turn a child’s attention to words, differing in their degree of elaboration. Labeling involves simply naming objects, characters, or actions in the book—for instance, saying

Labeling involves simply naming objects, characters, or actions in the book—for instance, saying “this is a forest”—while word explanations go further, offering definitions, descriptions, or elaborations, such as “when a place is filled with many trees, it’s called a forest.” The distinction matters because the two strategies are thought to support different aspects of vocabulary learning at different points in development. Labeling is concrete and direct, well suited to early word acquisition, whereas explanations become increasingly valuable as children’s language abilities grow and they can profit from richer, more elaborated talk about word meanings. Prior research suggests that parents naturally shift from one to the other as children mature, a sensitivity that reflects a broader tendency to tailor support to a child’s current abilities—what developmental theory describes as working within the child’s zone of proximal development.

What makes the new findings striking is that the difference between the two groups of families appeared in labeling but not in word explanations. Dyads who primarily read in their first language engaged in labeling more frequently than those who primarily read in their second language, yet the two groups did not differ in how often they provided word explanations. This asymmetry suggests that the choice of reading language shapes some forms of talk more than others. Naming an object in the moment may be easiest and most automatic in the language a family commands most comfortably, while offering an explanation—perhaps because it is a rarer, more effortful, and more deliberate move—may draw on resources that are less dependent on which language dominates the reading session.

The second major finding concerns code-switching, the alternating use of two languages within a single conversation or even a single utterance. Parents in the study code-switched more often during word explanations than during labeling, and this held true regardless of whether they primarily read in their first or second language. In other words, when parents moved from simply naming things to elaborating on what words mean, they became more likely to reach across their linguistic repertoire. This pattern offers a window into the functional, purposeful nature of bilingual speech: rather than mixing languages randomly, parents appear to deploy their second language precisely at moments of greatest communicative demand, when a child’s understanding may need extra support.

This kind of strategic language mixing is consistent with a growing body of research suggesting that bilingual parents use and combine their two languages for specific communicative purposes. For families raising children in a majority-language environment, the stakes of these choices are considerable. Parents from language-minority groups often must balance two competing goals: nurturing a minority heritage language at home while also preparing children for the language of formal schooling, which for most dual language learners is the majority language and their second language. Preschool entry frequently marks the beginning of systematic exposure to the majority language, which can place the minority language at risk of attrition. In the Norwegian context specifically, earlier survey work has found that parents report using their first language more when addressing their children than children do when addressing their parents—a pattern that underscores the parents’ central role as carriers of the heritage language.

Parents who wish to maintain their children’s proficiency in the heritage language are often aware of this responsibility and engage in deliberate practices to support it. The new study adds an important dimension to this picture by showing what those practices look like in the flow of actual reading interactions, rather than in parents’ retrospective reports. Because the data were collected naturalistically—families reading a wordless picture book at home, in their own space and on their own terms—the findings capture behavior that surveys and laboratory tasks may miss. Self-reports can be shaped by what parents believe they should do, and experimental designs can constrain the very behaviors of interest. Wordless picture books proved especially valuable here: unlike textbooks, which bilingual parents tend to read in the language of the text itself, wordless books free families to use either or both of their languages, revealing their natural preferences and habits.

The study also situates itself within a small but growing literature on how reading language relates to reading behavior. Previous work with English–French parents in Quebec found that parents reported reading more frequently and for longer stretches in their dominant language. Studies of Hispanic families in the United States have examined how parents’ preferred or primary language relates to their use of interactive and dialogic reading strategies, with mixed results—one study found that parents used strategies like asking questions and defining words most often when reading in their preferred language, particularly among those who preferred English, while another found no differences in dialogic strategies across language groups. By moving beyond self-report to direct observation, and by distinguishing between two specific forms of talk that differ in elaborateness, the Norwegian study refines the questions this literature can ask: it is not simply whether reading language matters, but which kinds of talk it shapes and which it leaves untouched.

The distinction between labeling and word explanations also connects to a broader research tradition on how parents draw children’s attention to words. Researchers have long noted that parental input during literacy activities varies from concrete to elaborated talk, and that word explanations in particular appear to be used infrequently during shared reading of textbooks. Studies examining whether mothers’ tendency to explain words relates to their educational level have produced mixed findings, suggesting that the drivers of elaborated talk are complex and not reducible to demographic variables. The present findings hint that the language situation of the family—specifically, which language anchors the reading session—may be one of those drivers, at least for the more automatic forms of word-focused talk.

There is also a developmental dimension worth emphasizing. As children grow more linguistically competent, parents may shift from labeling toward explanations, adapting their talk to match the child’s evolving abilities. The finding that code-switching clusters around explanations suggests that bilingual parents may treat their full linguistic repertoire as a flexible toolkit for this adaptation. When a simple label suffices, one language can carry the interaction; when a concept requires unpacking, the parent may reach for whichever language offers the clearest path to understanding, even mid-sentence. Such moment-to-moment decisions, repeated across countless reading sessions, could have cumulative consequences for how bilingual children build vocabulary in each of their languages.

The work builds on the research group’s own earlier qualitative observations of a small number of bilingual dyads, in which they noticed that parents who relied heavily on their first language actively worked to keep the conversation there—sometimes redirecting a child who used a second-language word by asking for the first-language equivalent. Scaling up to 151 dyads allowed the team to test whether the patterns glimpsed in that small sample held systematically across a larger, more diverse set of families. The confirmation that first-language-dominant dyads label more often, and that code-switching tracks explanatory talk across both groups, lends quantitative weight to what had been qualitative impressions.

For practitioners and policymakers, the findings carry practical implications. Shared reading programs aimed at bilingual families often encourage dialogic techniques, including asking questions and explaining word meanings. The evidence that parents naturally reserve their most elaborate, cross-linguistic talk for explanations suggests that such programs should acknowledge, rather than discourage, the strategic use of both languages. A parent who switches languages to clarify a word’s meaning is not undermining the heritage language project; they may be doing precisely what the interaction requires to keep the child engaged and comprehending. At the same time, the lower rate of labeling among second-language-dominant dyads raises questions about the vocabulary input children receive in each language during book time, and about whether families who read mainly in the majority language may be providing fewer of the quick, concrete naming moments that fuel early word learning in the heritage language.

More broadly, the study demonstrates the value of naturalistic observation for understanding bilingual family life. Language use at home is fluid, context-sensitive, and shaped by goals that parents may not articulate but reliably enact. By identifying systematic links between the language of reading, the type of talk, and the likelihood of code-switching, the research offers a template for future work: examining whether these patterns hold for other book types, older children, and other bilingual communities, and whether the talk patterns observed in reading predict children’s vocabulary growth over time in each of their two languages. In the meantime, the picture that emerges is one of quiet sophistication—parents orchestrating two languages with apparent effortlessness, in service of a child’s expanding understanding of words and worlds.

Subject of Research: Bilingualism

Article Title: Bilingual parent–child dyads' language use and code-switching during shared reading: exploring variations across labeling and word explanations

Article References: Nomat, D., Rydland, V., & Grøver, V. (2026). Bilingual parent–child dyads' language use and code-switching during shared reading: exploring variations across labeling and word explanations. Frontiers in Language Sciences, 5, Article 1934991. https://doi.org/10.3389/flang.2026.1934991

Image Credits: AI Generated

DOI: 10.3389/flang.2026.1934991

Keywords: code-switching, shared reading, bilingualism, labeling, word explanations, language development, parent-child interaction, scaffolding, dual language learners, Norway, Bilingual, parent

The authors frame their argument around a challenge statement that will resonate with anyone who has hired, or been, a first-year engineer. Undergraduate programs pour enormous energy into technical expertise, and rightly so, because employers still expect graduates to read schematics, interpret data, and understand regulatory constraints. But surveys of employer expectations and studies of the school-to-work transition consistently reveal a misalignment between what early-career engineers believe constitutes engineering expertise and what industry actually demands. Engineering work in practice is social, iterative, and frequently interrupted by failure. A design review can demolish months of effort in an afternoon. A regulatory submission can bounce back with a stack of deficiencies. Graduates trained primarily to execute well-defined academic assignments often find that the emotional and professional texture of industrial work was never part of the syllabus.

What makes the new framework notable is that it treats resiliency not as a personality trait students either possess or lack, but as a trainable professional competence that can be scaffolded across an entire undergraduate curriculum. The authors draw on a rich literature spanning resilience engineering in healthcare systems, psychological research on resilience enhancement in adults, and engineering education studies of design thinking. In the resilience engineering tradition, resilience is understood as the ability of a system to anticipate, absorb, and adapt to disturbances while continuing to function. Translated to the individual engineer, that means tolerating failed prototypes, revising assumptions under deadline pressure, and communicating honestly when a project slips. Psychological studies add that resilience can be strengthened through deliberate practice, reflective habits, and graduated exposure to manageable adversity, precisely the structure a well-designed curriculum can provide.

The framework begins in the first year, which the authors identify as the pivotal on-ramp. Rather than deferring design until a senior capstone, students encounter an introductory engineering design course immediately, one that normalizes iteration and failure from the outset. The pedagogical logic is grounded in decades of design education research showing that novice and expert engineers approach design problems in qualitatively different ways, and that early, structured exposure to open-ended problems accelerates the development of expert-like design processes. First-year students learn to define problems, brainstorm without attachment to their first idea, build rough prototypes, test them, and document what broke. Each cycle is a low-stakes rehearsal for the high-stakes failures of professional practice, and each recovery builds what education researchers call self-efficacy, the belief in one’s own capability to organize and execute the actions required to manage a task.

Self-efficacy is the study’s central measurable outcome, and the authors report that their data suggest introductory design courses, exposure to research opportunities, and team-based design projects led to improved self-efficacy and industry preparation among undergraduate biomedical engineers. This matters because self-efficacy is not mere confidence; it is a well-characterized psychological construct that predicts persistence in the face of difficulty. A student who has watched a fluidic device clog, diagnosed the failure, and rebuilt it is measurably more likely to persist through an industrial project crisis than a student who has only ever seen successful demonstrations. The framework deliberately engineers these experiences, sequencing them so that difficulty escalates as competence grows, following the scaffolded approach that has become standard in evidence-based curriculum design.

The second pillar of the framework is the transition from bench research at the student’s home institution to internships in industry, a pathway the authors describe as increasingly competitive. Internships have become the de facto gatekeepers of biomedical engineering careers, yet the skills that win a research assistantship in a university laboratory overlap only partially with those that secure a position at a medical device firm. The framework therefore treats the internship search itself as a design problem to be taught: students learn to translate academic project experience into industry-relevant narratives, to articulate teamwork and communication competencies alongside technical ones, and to present failed experiments as evidence of problem-solving rather than as liabilities. The authors outline a progression in which hands-on experience is layered deliberately, so that by the time students apply for industrial placements they have already practiced the professional behaviors those placements demand.

Teamwork receives particular emphasis, and for good reason. Biomedical products are built by cross-disciplinary teams of engineers, biologists, clinicians, regulatory specialists, and business strategists, and studies of workplace data on twenty-first-century STEM competencies consistently rank collaboration and communication among the most valued skills. The framework embeds team design projects throughout the curriculum, requiring students to navigate the same interpersonal frictions, unequal workloads, and conflicting technical opinions they will encounter professionally. The authors also address what education scholars call the hidden curriculum, the unwritten norms and professionalization pressures that shape how students learn to behave as engineers, arguing that making these norms explicit, including ethical reasoning and reflection, produces graduates who are not only skilled practitioners but thoughtful leaders committed to serving society with integrity and purpose.

Assessment in the framework relies on tools familiar to education researchers but underused in engineering programs. The authors describe the use of reflective instruments, including minute-paper style exercises in which students briefly record what they learned and where they struggled, allowing instructors to track the evolution of student perspectives across the academic journey from cornerstone to capstone. This longitudinal view is essential because research comparing freshman and senior design processes shows that growth in design thinking is gradual and uneven, and because studies of capstone courses alone suggest that a single culminating experience cannot compensate for four years without embedded professional skill development. By measuring self-efficacy repeatedly, the framework can identify which curricular interventions actually move the needle, rather than relying on end-of-program surveys that capture only a snapshot.

The broader significance of the work lies in its reframing of what biomedical engineering education owes its students. Accreditation bodies such as ABET have long required design content, and debates within the field have clarified why research experience cannot substitute for design experience in preparing students for clinical and industrial realities. But the Wheaton and Colorado-Denver framework goes further, arguing that design courses are not merely a vehicle for technical competence but the ideal training ground for professional resilience, because design is inherently a cycle of proposal, failure, revision, and persistence. In an industry where products take years to clear development and regulatory hurdles, and where the first prototype almost never ships, graduates who can metabolize failure quickly may be the single most valuable asset a company can hire.

The authors distill their aspiration into two postgraduation outcomes: that students will be prepared for professional endeavors and will demonstrate readiness for real-world challenges. Achieving those outcomes, they suggest, requires the kind of deliberate, scaffolded, four-year architecture their framework describes, one that starts with a first-year design course, weaves in research exposure and team projects, and treats resilience as a competency to be built rather than assumed. For faculty designing curricula, the message is that the path from bench to business runs not only through technical mastery but through the accumulated experience of failing, recovering, and trying again in an environment engineered to make that practice safe. For students, the message is equally direct: the setbacks you encounter in your first-year design course are not obstacles to your education. They are the education.

Subject of Research: Professional resilience training in undergraduate biomedical engineering curricula

Article Title: Bench to Business? Preparing Undergraduate Engineers for Careers in Biomedical Industry

Article References: Vazquez, K. J., & Benninger, R. K. P. (2026). Bench to Business? Preparing Undergraduate Engineers for Careers in Biomedical Industry. Biomedical Engineering Education. https://doi.org/10.1007/s43683-026-00253-6

Image Credits: AI Generated

DOI: 10.1007/s43683-026-00253-6

Keywords: biomedical engineering, engineering education, professional resilience, self-efficacy, design curriculum, undergraduate, internships, scaffolding, team-based learning, ABET accreditation, capstone design, industry readiness

The study arrives at a moment when classrooms around the world are being reimagined. For decades, educational theorists have argued that knowledge is something learners build rather than receive, a tradition stretching back to Lev Vygotsky’s work on the social origins of thought and Paulo Freire’s critique of “banking” models of education. Yet as Hocaoğlu and Ocak point out, co-constructivism is not a new idea in the literature so much as an under-specified one. The term appears across hundreds of papers on collaborative learning, teacher education, and instructional design, but its meaning often shifts from study to study. That vagueness matters, because without a clear operational definition, researchers cannot measure collaboration reliably, and teachers cannot know precisely what they are being asked to do differently.

To bring order to this conceptual sprawl, the researchers turned to the concept analysis method developed by L. O. Walker and K. C. Avant, a systematic procedure originally formulated for theory construction in nursing. The method proceeds through eight steps: selecting the concept, defining the aim of the analysis, identifying all known uses of the term, determining its defining attributes, constructing a model case along with alternative cases, identifying antecedents and consequences, and finally assembling relevant references. The approach forces the analyst to examine how a word is used both in dictionaries and in the scientific literature, gradually stripping away incidental associations until only the core attributes remain. It is a painstaking process, but one that yields definitions precise enough to support measurement instruments and operational research.

The literature base for the analysis comprised 62 articles selected for in-depth examination: 23 indexed in Web of Science, 20 in ERIC, and 19 in Scopus. These studies spanned an impressive range of educational settings, from elementary science classrooms and secondary mathematics lessons to medical education, teacher training programs, online discussion forums, and even community-based research with refugee populations. Some examined how children with autism and their teachers jointly built digital stories; others traced how nursing students co-created clinical leadership activities with their educators, or how YouTube comment sections became sites of informal knowledge co-construction. This breadth gave the authors confidence that the attributes they identified were not artifacts of any single context but genuinely constitutive features of the concept itself.

From this corpus, the researchers distilled five defining attributes of co-construction of learning: co-work, collective culture, culture sharing, co-construction of knowledge, and active involvement. Each attribute captures a distinct facet of the phenomenon. Co-work refers to the structured joint labor of teachers and learners, the concrete activities through which they pursue shared goals. Collective culture and culture sharing describe the social fabric that makes such labor possible, the norms of trust, mutual respect, and exchanged perspective that allow participants to treat one another as legitimate contributors. Co-construction of knowledge is the cognitive heart of the concept, the moment when ideas are jointly elaborated, challenged, and refined. Active involvement, finally, insists that all participants, not merely the most vocal, engage meaningfully in the process.

The analysis also clarified what must be in place before co-construction can occur, and what tends to follow when it does. Among the antecedents, the literature consistently highlighted meaningful teacher–student interactions and structured guidance, a finding that cuts against caricatures of constructivist classrooms as unstructured free-for-alls. Far from abandoning learners to discover everything alone, effective co-constructive teaching involves carefully scaffolded participation, in which the teacher’s role shifts from information transmitter to facilitator, guide, and fellow inquirer. The consequences identified were equally consequential for practice: deeper and more sustained learning, richer engagement, and a redefinition of the roles of both educators and learners across the various phases of curriculum implementation.

One of the study’s most practically valuable contributions is its potential to support the development of collaboration measurement instruments. Because the authors provide explicit definitions and defining characteristics of collaboration, researchers designing scales or observational coding schemes now have a validated conceptual foundation to work from. This matters for a field long plagued by measurement ambiguity. Prior attempts to quantify collaborative learning have often relied on proxies such as the frequency of group work or self-reported satisfaction, measures that may miss the substantive quality of joint thinking. A concept anchored in five well-defined attributes allows instrument designers to target those attributes directly, asking whether genuine co-work, shared culture, and active involvement are actually present in a given classroom.

The findings also resonate with a broader theoretical movement known as biocultural co-constructivism, which extends the logic of co-construction beyond the classroom to development itself. Scholars working in this tradition, including researchers studying goal orientations in social interaction and the neuropsychology of children in diverse cultural settings, argue that human development is jointly shaped by biology, culture, and individual agency in continuous reciprocal exchange. Within this larger frame, the classroom becomes a particularly visible arena for a process that is fundamental to human cognition: minds do not develop in isolation but through participation in socially organized activity. The Turkish study’s emphasis on socially situated learning thus connects a concrete pedagogical question to one of the deepest issues in developmental science.

For teachers, the study offers a reframing of everyday practice. Consider what the five attributes look like in a real lesson. A science teacher who asks students to predict, argue, and revise a model of seasonal change, rather than simply presenting the correct explanation, is engaging in model-based co-construction, a pattern documented in the analyzed literature on pre-service elementary teachers. Language instructors who treat grammar as a joint achievement, built through interaction rather than drilled in isolation, are doing the same thing in a different domain. Even assessment can be co-constructed: several studies in the corpus described students helping to build rubrics, generate exam questions, and participate in formative evaluation of their own writing. In each case, the defining features are the same, only the content changes.

The authors are careful to note the limits of their method. Concept analysis is a literature-based procedure; no new datasets were generated, and the study involved no human participants, which is why formal ethical approval was not required. The value of the work lies precisely in its synthesis: by systematically organizing how 62 studies used a contested term, Hocaoğlu and Ocak have given the educational community a shared vocabulary for one of its most important ideas. As schools continue to experiment with student-centered pedagogy, project-based learning, and technology-mediated collaboration, the study provides a conceptual compass. Learning, the analysis concludes, emerges not as passive knowledge acquisition but as an actively negotiated process, and recognizing that fact may be the first step toward classrooms designed around it.

Subject of Research: Concept analysis of co-construction of learning in educational settings

Article Title: Co-construction of learning: a walker & avant concept analysis

Article References: Hocaoğlu, N., & Ocak, G. (2026). Co-construction of learning: a walker &amp; avant concept analysis. Current Psychology, 45(17), Article 1477. https://doi.org/10.1007/s12144-026-10046-3

Image Credits: AI Generated

DOI: 10.1007/s12144-026-10046-3

Keywords: co-construction of learning, concept analysis, Walker and Avant, collaborative learning, constructivism, teacher-student interaction, educational psychology, classroom practice, knowledge building, Current Psychology, scaffolding, social learning

The study, led by Shahazwan Mat Yusoff, set out to answer a question that most previous Malaysian CEFR research had left unexplored. Earlier studies had documented that teachers generally hold positive views of the framework while struggling with training gaps, workload, materials, and student proficiency. What remained poorly understood was the cognitive and pedagogical process in between: how teachers actually make sense of the reform and translate it into daily decisions about tasks, scaffolding, and assessment. Drawing on the sensemaking tradition in education research, the authors argue that teachers do not simply receive policy messages and execute them. Instead, they interpret, select, modify, and sometimes quietly transform reform ideas according to their professional knowledge, beliefs, and the realities of their classrooms.

To capture this process, the researchers conducted semi-structured interviews with ten ESL teachers drawn from three government secondary schools, deliberately balanced between five novice teachers with fewer than five years of experience and five senior teachers with more than ten. Each interview lasted between thirty and forty-five minutes, and the transcripts were analysed thematically using both deductive codes derived from the research questions and inductive codes that emerged from the data itself. The team followed Braun and Clarke’s six-phase thematic analysis, strengthened by member checking with four participants, peer checking, and an audit trail of coding decisions. Ethical approval came from the Universiti Malaya Research Ethics Committee, and all participants gave written informed consent.

The first major finding is both encouraging and sobering. Teachers overwhelmingly viewed the CEFR as a useful, structured framework that brought clearer organisation to their lessons than the older KBSM syllabus. One novice teacher described it simply as more organised than what came before. Yet beneath this acceptance lay a striking unevenness in conceptual understanding. Some participants appeared to understand the CEFR mainly as a syllabus guide or textbook structure rather than as a proficiency-based framework for describing learner ability. A senior teacher captured the tension directly, noting that the real challenge was getting teachers to understand that the framework exists to describe what students can do. The distinction matters because a teacher can dutifully use CEFR-aligned materials while still interpreting the reform procedurally, missing its deeper purpose of informing judgements about learner progression and communicative task design.

The second finding shows how classroom realities became the crucible in which the reform was tested and reshaped. Teachers reported negotiating the CEFR against a cluster of practical constraints: limited instructional time, low and mixed learner proficiency, unfamiliar cultural content, and relentless examination pressure. Time emerged as one of the most immediate pressures, with one teacher admitting that restrictions forced the skipping of complementary skills within the prescribed lesson structure. Learner proficiency posed an even thornier problem. As one senior teacher explained, even in non-rural areas the number of students with low language capability and limited real-world knowledge was worrisome, and the end of ability streaming meant classrooms now contained a wide mix of levels. A single CEFR-aligned task could be trivial for some students and impenetrable for others, making implementation dependent on each teacher’s diagnostic judgement.

Cultural relevance added another layer of negotiation. Some teachers found that CEFR-aligned materials included topics, particularly around culture, that students simply could not relate to, dampening comprehension and participation. Meanwhile, the tension between communicative ideals and assessment demands produced genuine professional dilemmas. One senior teacher described having to decide whether to focus on a CEFR-based lesson or on the necessities of an upcoming examination, noting that the two are sometimes contradictory. The authors interpret these accounts not as evidence of teacher resistance but as contextual negotiation: teachers incorporated constraints into their pedagogical decision-making, simplifying tasks, localising content, and selectively implementing curriculum components. This supports the view, articulated in earlier scholarship, that Malaysia’s adoption of the CEFR is a case of global policy borrowing that must be reinterpreted within local educational and linguistic conditions.

The third finding, and perhaps the most provocative, concerns the difference between novice and senior teachers. Both groups recognised the need to adapt lessons to student needs, but the nature and specificity of their reported adaptations diverged sharply. Novice teachers tended to describe adaptation in broad, activity-focused terms: picking activities students could manage, adjusting question difficulty to proficiency levels, and relating lesson contexts to students’ lives. Senior teachers, by contrast, reported a considerably richer repertoire. One described first assessing the class’s proficiency profile, then designing activities to cater to the majority while scheduling personal intervention sessions for struggling learners. Another outlined differentiation strategies designed to challenge higher-level students while allowing lower-proficiency students to push themselves. Senior teachers’ accounts extended beyond task modification into scaffolding, motivation, targeted intervention, and support for low-anxiety communicative participation.

The authors are careful about what this comparison does and does not establish. Because the study relied on interviews rather than classroom observation, the findings describe differences in teachers’ reported adaptive repertoires, not directly observed differences in practice. The study does not claim that senior teachers are automatically better implementers of the CEFR. What it does suggest is that professional experience may shape the range and specificity of the strategies teachers can draw upon when negotiating learner diversity. This aligns with international research showing that experienced EFL teachers demonstrate stronger knowledge of scaffolding strategies, including the linguistic, affective, and metacognitive support that helps low-proficiency learners participate in communicative activities that would otherwise defeat them.

From these three strands, the researchers build a broader conceptual argument they call adaptive professionalism. Rather than treating adaptation as an add-on technique applied after the curriculum has been delivered, adaptive professionalism positions teachers as active interpreters of reform who use professional judgement to make global expectations workable in particular classrooms. The concept weaves together sensemaking theory, language teacher cognition research, and work on teacher agency, which holds that teachers’ capacity to act is shaped by the interaction of professional competence, structural conditions, and professional purposes. In this framing, the CEFR provides a common reference point, but it is teachers’ sensemaking and judgement that determine how its principles are understood and enacted. Implementation, in other words, is better understood as contextual negotiation than as policy transfer.

The practical implications are significant for Malaysia and for any education system importing an international framework. The findings suggest that professional development must move beyond familiarising teachers with CEFR terminology and descriptors. Teachers need opportunities to connect proficiency descriptors with actual learner performance, communicative task design, assessment decisions, and classroom progression, so that acceptance deepens into genuine conceptual understanding. They also need support in developing the adaptive repertoire that mixed-proficiency classrooms demand: scaffolding, differentiation, localisation of materials, and low-anxiety support for reluctant speakers. The authors acknowledge the study’s limitations, including its small sample from a single district and its reliance on self-reported practice, and they call for future research incorporating classroom observations, lesson artefacts, and student perspectives across rural and urban, high- and lower-performing settings. But the core message is already clear. Curriculum reform does not succeed on the strength of its design documents. It succeeds, or stalls, in the moment-to-moment interpretive work of teachers who must decide, lesson by lesson, what a global framework means for the specific students sitting in front of them.

Subject of Research: Teacher sensemaking and adaptive enactment of the CEFR-aligned English curriculum in Malaysian secondary ESL classrooms

Article Title: Teacher sensemaking and adaptive enactment of curriculum reform based on the CEFR in Malaysian secondary ESL classrooms

Article References: Mat Yusoff, S., Mohamad Marzaini, A. F., Ahamat, M. I., Abdullah, M. R., Seyapalan, E. D., & Ali Huddin, M. R. (2026). Teacher sensemaking and adaptive enactment of curriculum reform based on the CEFR in Malaysian secondary ESL classrooms. Discover Education, 5(1), Article 1152. https://doi.org/10.1007/s44217-026-02305-z

Image Credits: AI Generated

DOI: 10.1007/s44217-026-02305-z

Keywords: CEFR, ESL teachers, curriculum reform, teacher sensemaking, Malaysia, qualitative case study, scaffolding, differentiation, language policy, communicative language teaching, novice and senior teachers, adaptive professionalism

The stakes are high in a country where spoken English opens doors to social, academic, and professional opportunity, yet where proficiency remains stubbornly low. Pakistan ranked 67th among 116 countries in the Education First English Proficiency Report, and earlier research by the same team found that 88 percent of secondary school students performed at only the most basic speaking levels, classified as A1 or A2 on the Common European Framework of Reference (CEFR). Just 12 percent reached the pre-intermediate B1 level, and not a single student achieved advanced proficiency. The reason, the researchers argue, is structural: Pakistani public schools emphasize rote memorization and written exercises, leaving students with years of English instruction but almost no meaningful oral practice.

The study was grounded in Vygotsky’s Sociocultural Theory, which holds that learners advance through interaction with a More Knowledgeable Other within their Zone of Proximal Development, the space between what a learner can do alone and what they can achieve with guidance. In this framing, ChatGPT served as a scaffold: an endlessly patient conversational partner that adjusted its questions to the student’s level, prompted elaboration, and corrected errors in real time. After each individual practice session, students joined collaborative classroom activities such as role plays, pair work, and group discussions, allowing them to consolidate what the AI practice had unlocked.

The intervention itself was built with the ADDIE instructional design model and aligned with the Grade 9 English syllabus. Each week began with a five-minute teacher briefing, followed by sixty minutes of ChatGPT-assisted speaking practice in the school’s IT laboratory, and closed with twenty-five minutes of collaborative learning. Students used standardized prompts to steer the conversation. In one documented exchange on the topic of child labor, a student first asked ChatGPT to act as an English tutor and provide a word bank of useful vocabulary, then asked factual questions about the topic, and finally instructed the AI to act as a speaking partner, asking questions one at a time and giving brief feedback after each answer. Crucially, students kept the same conversation thread throughout the twelve weeks, so the AI retained the history of prior interactions and could build on them.

Measuring the outcome required a rigorous assessment framework. The researchers developed a Spoken English Proficiency Test modeled loosely on the IELTS structure, comprising a warm-up conversation, a four-to-five-minute monologue, and a five-to-six-minute dialogue. Each task was scored against CEFR descriptors across five dimensions: range, the breadth of vocabulary and expression; accuracy, grammatical correctness; fluency, the smoothness of delivery; interaction, the ability to engage a conversational partner; and coherence, the logical organization of speech. Each component was scored from 1 to 6, corresponding to CEFR levels A1 through C2, for a maximum total of 60 points. The instrument showed exceptional internal consistency, with a Cronbach’s alpha of 0.964.

In a novel twist, the assessment itself was conducted by ChatGPT. The researchers uploaded the CEFR scale, the marking scheme, and the test materials, then prompted the AI to act as a CEFR English-speaking examiner. Students spoke directly to ChatGPT in Voice Mode while the researcher supervised, and the AI generated real-time transcripts and scored the speaking components. Human language experts, blinded to group allocation, oversaw quality control and independently scored the interaction component of the monologue task, since ChatGPT could not evaluate non-verbal engagement such as facial expressions and gestures. This human-AI collaborative approach aimed to combine standardized, consistent scoring with expert judgment.

The results were unambiguous. Before the intervention, the two groups were statistically equivalent, with all students clustered in the basic A1-A2 range. After twelve weeks, the experimental group’s mean overall score reached 36.56, compared with 17.60 for the control group, a difference the Mann-Whitney U test confirmed as highly significant, with U = 1.50, z = -6.08, and p < .001. The effect size of r = .86 is unusually large for an educational intervention. Perhaps more telling is the raw gain: the ChatGPT group improved by an average of 20.44 points, while the control group gained just 1.04. On the CEFR scale, experimental students progressed from Basic User to Independent User levels, with one learner reaching C1, the threshold of Proficient User.

The improvements held across every measured dimension and both task types. In monologue performance, where students spoke independently on topics such as their dream job, the experimental group significantly outperformed the control group on range, accuracy, fluency, interaction, and coherence, all at p < .001. The same pattern emerged in dialogue tasks, where students conversed on everyday topics such as weekend plans. This across-the-board pattern matters because previous studies of ChatGPT in language education tended to focus on motivation, confidence, or overall proficiency; this study is among the first to isolate and document gains in the specific qualitative dimensions that define spoken competence under an internationally recognized framework.

The findings align with a growing body of international evidence. Studies in other contexts have reported that interactions with ChatGPT improve oral speaking performance, that consistent engagement with AI-assisted speaking activities produces greater gains than sporadic use, and that twelve-week chatbot interventions help learners outperform traditionally taught peers. Researchers have also documented improvements in vocabulary use, grammatical accuracy, and fluency from repeated AI-supported practice. Yet the literature carries caveats: systematic reviews warn about ethical concerns, uneven quality of AI-generated feedback, and the fact that chatbots cannot fully replicate human interaction. Effectiveness, the evidence suggests, depends less on the technology itself than on pedagogical integration, teacher guidance, and instructional design, all of which were central to the Pakistani module.

The authors are careful about the limits of their claims. The sample comprised fifty female students from a single urban public school, a homogeneous group in age, socio-economic background, and educational history, which may have reduced variability and inflated effect sizes. The twelve-week duration cannot speak to long-term retention, and the researchers recommend delayed post-tests, larger and more diverse samples, and mixed-methods designs to capture how the learning actually happens. The theoretical interpretation, that ChatGPT provided scaffolding within students’ zones of proximal development, remains a plausible explanation rather than a directly demonstrated mechanism. Still, the practical implications are hard to ignore. For classrooms where one teacher faces dozens of students and individual speaking practice is a logistical impossibility, an AI conversational partner offers a scalable supplement, and the study suggests that policymakers weighing digital infrastructure and teacher training for AI-supported language education now have experimental evidence to inform those decisions.

Subject of Research: Effectiveness of a ChatGPT-assisted learning module for improving English-speaking proficiency among secondary school students

Article Title: Integration of ChatGPT to enhance English speaking skills among secondary school students in Pakistan

Article References: Nageen, S., Khawaja, H. U. H., Sarwar, M., Moin, M., Imran, Z., & Jabeen, S. (2026). Integration of ChatGPT to enhance English speaking skills among secondary school students in Pakistan. Discover Education, 5(1), Article 1137. https://doi.org/10.1007/s44217-026-02278-z

Image Credits: AI Generated

DOI: 10.1007/s44217-026-02278-z

Keywords: ChatGPT, artificial intelligence, English speaking proficiency, CEFR, Pakistan, secondary education, language learning, experimental study, educational technology, Vygotsky, scaffolding, randomized controlled trial

Metacognition, in the framework the researchers adopted, covers both knowledge about one’s own cognition and the regulation of cognitive activity through planning, monitoring, evaluation, and strategic control. Digital environments are natural arenas for these processes because they flood learners with choices—multiple pathways, abundant resources, automated feedback—and demand that learners judge whether their strategies are working. Technologies can be designed to scaffold regulation directly: prompts can nudge goal setting, dashboards can make progress visible, learning diaries can structure reflection, and feedback systems can highlight the gap between intended and actual performance. With the arrival of intelligent tutoring systems, learning analytics, and now generative AI capable of open-ended explanation and critique, the repertoire of possible support has expanded dramatically. But more capable systems can also perform regulatory work on the learner’s behalf, raising a question that runs through the entire review: does the technology strengthen independent regulation, or quietly absorb it?

To answer such questions rigorously, the team followed PRISMA 2020 reporting standards and searched five major databases—Web of Science, Scopus, ERIC, IEEE Xplore, and PsycINFO—completing searches on 9 July 2026 with no lower publication-year limit. The initial sweep returned 5,326 records. After removing 2,532 duplicates, 2,794 unique records underwent title-and-abstract screening, and 168 reports were assessed in full text. Thirty-three studies ultimately met all eligibility criteria: each had to involve students in formal K–12 or post-secondary settings, compare a technology-supported intervention against a control or business-as-usual condition, and report a quantitative metacognitive outcome from which a standardised effect size could be derived. Two researchers screened and coded everything independently, achieving Cohen’s kappa values between 0.84 and 0.94 on categorical variables and intraclass correlations above 0.94 on continuous quantities—agreement figures that signal unusually disciplined extraction.

The statistical engine of the study was a three-level random-effects meta-analysis, a model chosen because the 33 studies contributed 94 effect sizes and multiple effects from the same study cannot be treated as independent. Level 1 of the model captured sampling variance, Level 2 captured heterogeneity among effect sizes within studies, and Level 3 captured heterogeneity between studies. The pooled estimate of g = 0.653 was highly significant, but the heterogeneity was substantial: total I² reached 72.7 percent, and virtually all of the excess variance sat at the between-study level, with the within-study component estimated at effectively zero. A 95 percent prediction interval stretching from roughly −0.02 to 1.32 drives the point home. In a comparable future implementation, the true effect could plausibly be negligible—or large. The average, in other words, describes a scattered landscape rather than a uniform intervention effect.

Perhaps the most revealing part of the analysis is a descriptive map of what the interventions actually did. Evaluation and reflection were explicitly supported in 29 of 33 studies, or 87.9 percent; monitoring and strategic control each appeared in 26 studies, or 78.8 percent; but planning was represented in only 19 studies, or 57.6 percent. The pattern suggests that most digital interventions intervene after learning is under way—prompting learners to inspect progress, self-assess, and revise—while far fewer help them set goals and select strategies before starting. Because goals and task analysis provide the foundation on which monitoring and evaluation depend, the authors argue that future designs should attend to the full regulatory cycle rather than its back half.

The outcome evidence was even more lopsided. Metacognitive regulation accounted for 77 of the 94 effect sizes, or 81.9 percent, while metacognitive knowledge contributed just 11 effects and composite measures six. More striking still, 90 of 94 effects—95.7 percent—rested on self-report instruments. Performance-based, judgement-based, or calibration indicators produced only three effects, trace or process data contributed one, and not a single study used researcher-rated task products. This means the pooled estimate primarily summarises changes in what learners say about their own planning, monitoring, and reflection, not independently demonstrated regulatory skill. Self-reports are legitimate data, and learners’ perceptions of their strategy use can genuinely change in response to instruction, but perceived regulation is not equivalent to accurate or effective regulation enacted during performance—a distinction long emphasised in the metacognition literature.

The exploratory moderator analyses generated the study’s most newsworthy, and most qualified, findings. Three implementation features cleared the conventional unadjusted significance threshold: interventions with explicit technology-use guidance outperformed unguided ones (g = 0.760 versus 0.479), immediate feedback beat delayed feedback (g = 0.744 versus 0.408), and system-directed learner control beat learner-directed control (g = 0.780 versus 0.499). But none of these survived Benjamini–Hochberg false-discovery-rate correction across the eight moderator tests, with each adjusted p value landing at 0.095. Crucially, the three features overlapped heavily in practice—14 of the 20 guided studies also provided immediate feedback—so their apparent effects may reflect a bundled instructional style rather than three separable design principles. The authors explicitly frame these patterns as hypothesis-generating rather than as established moderator effects.

Technology type, notably, did not significantly moderate outcomes. Generative-AI interventions showed the largest descriptive subgroup estimate at g = 0.831, compared with 0.643 for non-AI technologies and 0.346 for non-generative AI, but those subgroups contained only nine and four studies respectively, drawn from heterogeneous systems. The findings therefore do not support any claim that AI, or generative AI in particular, is intrinsically superior for building metacognition. Educational level, subject area, and intervention duration also failed to explain meaningful variation. The robustness of the headline result, by contrast, was impressive: cluster-robust inference reproduced the same point estimate, imposing assumed within-study correlations from 0.30 to 0.90 shifted the pooled estimate only between 0.640 and 0.651, leave-one-study-out estimates ranged from 0.625 to 0.675, and excluding the two studies rated at serious risk of bias left the association intact at g = 0.624.

The risk-of-bias picture tempers enthusiasm further. Among the 33 studies, only nine used randomised controlled designs; 22 quasi-experimental studies were judged at moderate risk under ROBINS-I, two at serious risk, and no study at all received a low overall risk rating. A cluster-robust Egger-type precision test found no statistically detectable evidence of small-study effects, though the authors interpret this as an absence of evidence rather than evidence that publication bias is absent. These caveats, combined with the self-report dominance, mean the pooled effect is best read as a positive average association in reported metacognitive regulation rather than proof of uniformly improved metacognitive capability.

The practical upshot is a design philosophy rather than a shopping list. The authors argue that technology selection should begin with the regulatory problem to be solved—goal setting, monitoring, evaluation, or strategy adaptation—rather than with the technological label, and that AI systems in particular should be judged by which regulatory decisions remain with the learner and how support is faded over time. For researchers, the priorities are larger controlled studies, transparent descriptions of what each design feature is meant to elicit, and multimethod assessment combining self-reports with calibration judgements, performance indicators, and behavioural traces. As generative AI spreads through classrooms, the difference between a tool that cultivates independent thinkers and one that induces what other researchers have called metacognitive laziness may depend on exactly those choices. This synthesis gives educators a genuine, if qualified, reason for optimism—and a precise agenda for finding out what actually works.

Subject of Research: Effects of technology-supported educational interventions, including AI tools, on learners' metacognition

Article Title: Technology-Supported Interventions and Learners’ Metacognition: A Three-Level Meta-Analysis

Article References: Yuan, Y., & Du, Y. (2026). Technology-Supported Interventions and Learners’ Metacognition: A Three-Level Meta-Analysis. Educational Psychology Review, 38(1), Article 130. https://doi.org/10.1007/s10648-026-10227-3

Image Credits: AI Generated

DOI: 10.1007/s10648-026-10227-3

Keywords: metacognition, educational technology, meta-analysis, self-regulated learning, generative AI, artificial intelligence in education, learning analytics, feedback, scaffolding, self-report measures, educational psychology, learner control

The study, published in Discover Education, combined a close theoretical reading of the official textbooks with empirical testing of real students. The researchers analyzed both the teacher and student editions of the mandated textbooks using the Anthropological Theory of the Didactic, a framework developed by the French didactician Yves Chevallard that treats mathematical knowledge as a set of organized practices. Central to this framework is the concept of praxeology, the pairing of tasks and techniques with the explanations and theories that justify them. A well-designed textbook, in this view, does not merely show students how to graph an inequality; it explains why the shaded region represents an infinite set of solutions and why two overlapping regions must both be satisfied simultaneously. The Indonesian textbooks, the analysis found, repeatedly present the how while omitting the why.

That imbalance has a technical name in the literature: a praxeological imbalance, in which the practical block of tasks and techniques crowds out the logos block of justification and theory. The researchers documented concrete examples. In one opening problem, students weigh apples and oranges on a scale and are asked to build a system of inequalities through trial and error. The task is engaging, but it never explicitly instructs students to convert kilograms to ounces, a hidden technical demand that derails learners who have not mastered unit conversion. The completed mathematical model is then printed directly beside the problem, inviting imitation rather than independent reasoning. Elsewhere, the book leaps from single inequalities to full systems without scaffolding the crucial idea that a system’s solution is the intersection of two half-planes, not simply the area below each line.

To test whether these theoretical weaknesses translate into real difficulties, the team administered a diagnostic assessment of eight open-ended tasks to 80 students at a public vocational school in Banten, Indonesia, and then conducted in-depth interviews with 20 of them, selected purposively for showing distinctive error patterns. The tasks were deliberately mapped to the obstacle categories predicted by the textbook analysis: four targeted epistemological obstacles involving inequality symbols, solution sets and half-planes, three targeted didactical obstacles arising from instructional design, and one addressed ontogenic obstacles in applying the concepts to real-world contexts. The design meant that each error a student made could be traced back to a specific feature of how the material had been presented.

The results were remarkably consistent with the predictions. Half of the students, 50 percent, could draw boundary lines but could not interpret what the inequality symbols meant graphically, leaving them unable to decide which side of the line to shade. One student told the researchers, I have drawn the lines, but I don’t know whether to shade the upper or lower part. Twenty percent failed to identify the intersection of two regions as the solution set, instead shading everything below both lines; one admitted, I didn’t know I had to look for the intersection, I thought it was enough to shade below the lines. Another 30 percent chose test points randomly from the graph without substituting them into the inequalities, revealing a disconnect between graphical and symbolic representations that echoes Raymond Duval’s classic work on the cognitive difficulty of converting between mathematical registers.

The researchers also catalogued ontogenic obstacles, those rooted in students’ own developmental readiness. Conceptual gaps were the most common, affecting 55 percent of participants, who lacked command of prerequisites such as plotting lines and finding axis intercepts. Instrumental obstacles, meaning weak technical skills like substitution and arithmetic, appeared in 30 percent; one student wrote that 3 plus 3 equals 7 while checking a point against the inequality x plus y is at most 7. Psychological obstacles, including math anxiety and fear of failure, affected 15 percent, with students reporting they did not dare to try because they were afraid of being wrong. These affective barriers, the authors note, consume working memory resources and compound the conceptual difficulties, a finding consistent with the broader literature on cognitive load and mathematics anxiety.

Perhaps the most consequential category was the didactical obstacle, the kind produced by instruction itself. Forty-five percent of students shaded regions by pure imitation of textbook patterns, with one explaining, I shade the area below the line because that’s how it is in the book, but I’m not sure if it’s correct. Thirty-five percent had never been taught to seek the intersection of two constraints, and 20 percent showed no habit of algebraic validation at all, selecting points visually because, as one put it, the teacher usually never asks to check points. The researchers interpret these behaviors as products of an implicit didactical contract in which students see mathematics as a sequence of authoritative steps to be copied rather than a logical system requiring justification. When the textbook provides worked templates but no validation prompts, students learn that verification is not part of the job.

The team argues that these findings establish more than a correlation. By triangulating the textbook task layouts, the diagnostic error patterns and the interview testimonies, they built a mapping matrix linking specific textbook features to specific student errors. The high rate of shading mistakes corresponds to the books’ emphasis on mechanical table completion and the placement of finished models next to problems, which lowers cognitive demand and invites copying. The intersection and validation errors correspond to the textbooks’ silence on those very procedures. In the researchers’ framing, the obstacles are structurally embedded in the instructional material, not merely individual deficits, which means they can be anticipated and designed away.

The proposed remedies are concrete. The authors recommend a three-phase scaffolding model in which students first explore a contextual problem with guiding questions, then compare their reasoning with a partially completed model, and only finally see a fully worked example accompanied by explicit justification of each inequality. They call for standardized validation prompts after every graphing task, prerequisite activation boxes, and what-if prompts that ask how changing a strict inequality to an inclusive one alters the boundary. Comparative context strengthens the case: studies of Chinese textbooks by Liping Fan and Yan Zhu found more systematic scaffolding and explicit linking of models to tasks, suggesting the Indonesian materials lag behind international best practice. The researchers also urge a formal praxeological review of all government-published textbooks, so that every technique is paired with its conceptual rationale, and recommend that teachers supplement the books with think-aloud protocols and activities that make students justify their shading choices aloud.

The study’s limitations are acknowledged: it examined a single vocational school population and did not test an alternative textbook design longitudinally, so generalization requires caution. Yet its central message travels well beyond one Indonesian classroom. Textbooks are the primary, and often the only, reference in many mathematics classrooms worldwide, and any systematic weakness in their design propagates quietly through millions of learners. Systems of linear inequalities matter because they model budget constraints in economics and resource limits in engineering, and students who master them only mechanically tend to falter in linear programming, optimization and calculus. If the obstacles are built into the pages, the researchers conclude, then the fix belongs there too, in a deliberate redesign that treats conceptual understanding, solution validation and reflection not as extras but as the core of what a mathematics textbook is for.

Subject of Research: The impact of mathematics textbook design on learning obstacles and student performance in systems of linear inequalities in two variables

Article Title: Assessing the impact of mathematics textbooks on learning obstacles and student performance in systems of linear inequalities in two variables

Article References: Nasrullah, A., Suryadi, D., Prabawanto, S., Hendriyanto, A., Muhaimin, L. H., & Dewi, D. A. (2026). Assessing the impact of mathematics textbooks on learning obstacles and student performance in systems of linear inequalities in two variables. Discover Education, 5(1), Article 1036. https://doi.org/10.1007/s44217-026-02064-x

Image Credits: AI Generated

DOI: 10.1007/s44217-026-02064-x

Keywords: mathematics education, textbook analysis, linear inequalities, learning obstacles, didactic transposition, praxeology, Merdeka Curriculum, algebra learning, cognitive load, math anxiety, scaffolding, Indonesia

The most eye-catching result concerned compound interest, one of the foundational concepts of personal finance and a topic that decades of research show people worldwide understand poorly. On an objective multiple-choice item requiring students to convert a monthly interest rate into its annual equivalent, the correct-response rate exploded from just 3.1 percent before the learning period to 81.2 percent afterward, a jump of more than 78 percentage points. Performance also improved on items covering simple interest, early savings growth, and budgeting, while an exchange-rate question remained flat at 65.6 percent. Because the same five objective items were administered twice, the authors caution that practice effects could account for part of the improvement, but the sheer scale of the compound-interest gain suggests that something more than mere familiarity with the test format was at work.

The chatbot’s design is arguably as interesting as its results. Rather than giving the model free rein, the researchers built it as a web application on the GPT-4o API and governed its behavior entirely through a structured system prompt, without retrieval-augmented generation or an external knowledge base. The system prompt encoded four pedagogical principles. First, scope restriction: the bot could discuss only financial literacy topics within the six measurement domains and was explicitly forbidden from giving personalized investment advice, real-time market data, or speculative forecasts. Second, pedagogical scaffolding: it was configured as a step-by-step learning assistant that answered every question through a fixed sequence of conceptual explanation, worked example, and reinforcement question. Third, safety redirects: questions about current tax law or specific investment products were referred to authoritative institutions such as Türkiye’s Capital Markets Board and Revenue Administration. Fourth, linguistic accessibility: all responses were delivered in plain Turkish with technical terms defined inline.

A documented example exchange shows the scaffold in action. When a student asked what compound interest means, the chatbot first explained the concept, noting that it means earning interest on both the principal and previously accumulated interest. It then offered a worked example: depositing 1,000 Turkish lira at 10 percent annual interest compounded monthly would yield roughly 1,105 lira after one year rather than 1,100. Finally, it posed a reinforcement question, asking the student to calculate what 500 lira would become after two years at the same rate. This explain-exemplify-elicit structure was written into the system prompt as the default response pattern, translating three decades of learning science into conversational practice.

That structure draws deliberately on three theoretical pillars. Constructivist learning theory, rooted in the work of Piaget and Vygotsky, holds that learners build knowledge actively rather than passively absorbing it, and the chatbot’s scaffolding was designed to operate within the learner’s zone of proximal development, the space between what a student can do alone and what they can achieve with guidance. The immediate feedback model, articulated by Hattie and Timperley, supplies the rationale for rapid, task-relevant corrective responses, which matter enormously in a domain like finance where a single misunderstood calculation can cascade. Self-determination theory, from Ryan and Deci, motivated the decision to let students direct their own questioning during the learning week, supporting the psychological needs for autonomy and perceived competence. The authors are careful to frame these theories as design rationales rather than tested causal mechanisms, since the study did not measure how students actually used the scaffolding.

The measurement itself was unusually detailed. The primary instrument comprised 32 factual financial statements, eleven of them deliberately worded as misconceptions, which students rated on a six-point agreement scale before being keyed against correct answers. Items were organized into six subscales: Basic Economics and Finance, Individual Banking, Retirement and Insurance, Financial Statements, Investment, and Tax and Legislation. Five objective multiple-choice questions on exchange rates and financial mathematics served as a complementary performance check. The researchers chose paired-samples t-tests for subscales whose difference scores passed Shapiro-Wilk normality tests and Wilcoxon signed-rank tests for those that did not, reporting bootstrap confidence intervals for the rank-based effects. The largest subscale gains appeared in Financial Statements, with a mean increase of 1.08 points and an effect size of 0.95, and Retirement and Insurance, which rose 0.83 points with an effect of 0.91, while Individual Banking showed the smallest change at 0.34 points, plausibly because students started from a high baseline of familiarity with internet banking and credit cards.

Not every data point moved in the expected direction, and one reversal is particularly instructive. Agreement with the false statement that inflation in Türkiye is below 10 percent actually rose during the study, producing the largest single-item change in the entire instrument once the item was reverse-coded, a decline of 1.50 points. Thirteen students answered this item less accurately at post-test, sixteen were unchanged, and only three improved. The authors offer a telling explanation: the chatbot was deliberately configured without access to real-time market data and instructed to redirect questions about current economic conditions to authoritative sources, so it could explain what inflation is but could not state what the current rate actually is. The decline on this item is therefore consistent with the system’s scoping rather than contradicting it, though the researchers acknowledge that students may also have generalized from worked examples that used round illustrative rates, or that the item itself may simply be unstable.

The study’s honesty about its own limits is notable. With no control or comparison group, the observed changes cannot be causally attributed to the chatbot; testing effects, maturation, concurrent coursework, novelty, and regression to the mean all remain plausible contributors. Interaction logs were never systematically recorded, so session duration, question counts, and topic coverage cannot be reconstructed. Several subscales showed weak or even negative pre-test internal consistency, with Cronbach’s alpha values as low as minus 0.22 for the Investment subscale, meaning dimension-specific findings must be treated as exploratory. The agreement-scale format also conflates knowledge with confidence, and the small volunteer sample of 32 students from a single university constrains generalizability. The authors explicitly position the work as a design and feasibility record rather than evidence of effectiveness, and they lay out a roadmap for follow-up research: randomized controlled trials, verified knowledge sources such as retrieval-augmented generation for regulation-sensitive content, systematic privacy-aware logging, validated instruments with established factor structures, delayed post-tests to separate practice effects from retained learning, and outcome measures that capture actual financial behavior rather than agreement-rated knowledge alone.

Even with those caveats, the findings land at a moment when both educators and technologists are searching for scalable answers to a persistent problem. Financial literacy is strongly linked to retirement planning, wealth accumulation, and resistance to fraud, yet surveys consistently show that even basic understanding of compound interest, inflation, and risk diversification remains shockingly low worldwide. University students face a particularly acute transition, juggling credit cards, student loans, and housing costs as they approach financial independence, while traditional instruction cannot offer continuous, individualized explanation on demand. A scaffolded chatbot can, at low marginal cost. The Bayburt team’s work suggests that the recipe matters as much as the technology: constrain the scope, forbid personalized advice, route regulation-sensitive questions to authoritative institutions, and build the pedagogy directly into the prompt. Meta-analyses of educational chatbots report modest pooled effects, with a 2026 synthesis of 35 experimental studies finding a moderate effect of ChatGPT on learning outcomes, so the large uncontrolled changes observed here demand cautious interpretation. But as a blueprint for how large language models might be tamed into patient, safe, and pedagogically disciplined financial tutors, this pilot offers one of the clearest demonstrations yet, and a compelling reason for controlled trials to follow.

Subject of Research: A pilot pre-post study of a scaffolded GPT-4o AI chatbot for financial literacy education among university students

Article Title: A pilot pre–post study of a scaffolded AI chatbot for financial literacy education among university students

Article References: Ayberkin, D., & Yirik, E. (2026). A pilot pre–post study of a scaffolded AI chatbot for financial literacy education among university students. Discover Education, 5(1), Article 1068. https://doi.org/10.1007/s44217-026-02226-x

Image Credits: AI Generated

DOI: 10.1007/s44217-026-02226-x

Keywords: financial literacy, AI chatbot, GPT-4o, large language models, university students, educational technology, scaffolding, pre-post study, compound interest, Türkiye, higher education, pilot study