Psychology by BogdanovaThe Encyclopedia of Psychology for the Artificial Era

RESEARCH DOSSIER

Digital, Online, and Blended Learning Environments

Abstract. Digital, online, and blended learning environments are organized educational ecologies in which technologies shape access to instruction, the temporal organization of participation, social and instructional interaction, feedback, assessment, and the conditions under which knowledge is retained and transferred. This monograph develops a critical scientific synthesis of research on distance education, learning sciences, cognitive psychology, self-regulation, multimedia processing, social presence, technology-mediated pedagogy, educational equity, and emerging artificial tutoring. It separates delivery medium from instructional treatment, differentiates synchronous, asynchronous, hybrid, and blended arrangements, reconstructs major theoretical and methodological developments, and evaluates representative comparative trials, meta-analyses, and primary experiments. The synthesis establishes that educational effectiveness depends on interactions among learners, teachers, content, institutions, task design, and access conditions. The evidence supports well-designed interaction, feedback, and active cognitive work, while showing substantial heterogeneity, confounding in many blended-learning comparisons, gaps in long-term transfer studies, and limits to causal interpretations of platform engagement indicators. The final chapters address emergency remote schooling, equity, multimodal access, responsible artificial assistance, and a research agenda for durable, independently demonstrated learning.

Keywords: online learning; blended learning; hybrid instruction; digital learning environments; distance education; synchronous learning; asynchronous learning; educational psychology; community of inquiry; self-regulated learning; cognitive load; multimedia learning; formative assessment; social presence; digital equity; learning analytics; artificial tutoring; learning transfer.

Introduction and Research Object

The scientific object of digital, online, and blended learning environments is the organization of learning activity under technologically mediated conditions. A learning environment comprises instructional arrangements, artifacts, relationships, representations, rules of participation, and opportunities for feedback. A digital environment changes how these elements are distributed across space and time, how readily a learner can return to instructional content, and what forms of assistance or interruption accompany activity. The central research question concerns how those arrangements affect understanding, skill acquisition, retention, transfer, motivation, participation, and inclusion.

Educational technology is frequently evaluated as though the presence of a device, learning-management system, video recording, or videoconference constituted a complete educational intervention. Such a description omits instructional duration, the quality of explanations, the opportunities for practice, assessment incentives, instructor contact, accessibility, and the learner's prior knowledge. The same platform can support inquiry, memorization, unstructured browsing, deliberate practice, or disengagement. Comparative research consequently requires a description of the educational treatment rather than a simple binary variable indicating whether teaching occurred online.

The monograph considers formal primary, secondary, postsecondary, professional, and adult learning where evidence is relevant, while recognizing that direct comparisons are concentrated disproportionately in higher education. Its analytic focus is the psychology of learning and educational participation. Software engineering, broadband infrastructure, and institutional finance enter the analysis where they explain the opportunities available to learners. The associated but separate study of fully artificial learning belongs to its own research object. Human learning through artificial assistance is addressed here only insofar as the digital educational environment shapes human performance.

The present work is an evidence-informed critical narrative synthesis. It traces identified foundational studies, experimental contrasts, systematic reviews, meta-analyses, and official research reports, and distinguishes their designs and interpretations. It does not claim a preregistered search strategy, exhaustive evidence capture, an original participant sample, or a newly calculated pooled effect size. The systematic review by Martin, F., Sun, T., & Westine, C. D.(2020) is instructive because its map of research from 2009 to 2018 shows how strongly the field has emphasized learner engagement and higher-education contexts, leaving organizational and cross-context evidence less developed.

Definitions and Boundaries of Digital Learning

Digital learning refers broadly to learning activity facilitated, represented, assessed, or organized through digital technology. Online learning is the subset in which networked environments mediate material instructional processes, communication, or assessment. Distance education is older and broader: spatial separation can be mediated by correspondence, broadcasting, or other technologies. Synchronous online learning requires participants to coordinate in real time; asynchronous learning permits material engagement and interaction at different times. The distinction concerns temporal coordination, not intrinsic educational quality.

Blended learning combines face-to-face and digitally mediated learning in an intentionally integrated course or program. Its scientific significance lies in how the components divide pedagogical labor: initial explanation, active problem solving, communication, feedback, rehearsal, and assessment may take place through different modalities. A course with a classroom meeting and an unrelated repository of downloadable files is technologically supplemented, but its effectiveness cannot be inferred from research on deliberately integrated blends. Hybrid is used inconsistently across institutions: sometimes it is a synonym for blended learning, and sometimes it specifies alternating attendance or simultaneously serving remote and present students. Scientific reporting must state which arrangement actually occurred.

An educational environment includes temporal flexibility, interaction architecture, feedback availability, representational design, assessment practices, accessibility, instructor support, and institutional rules. These dimensions can vary independently. A fully online course with guided practice, frequent feedback, and structured peer interaction may resemble a well-designed classroom educationally more than a passive online lecture resembles an active online seminar. This dimensional account is consistent with the meta-analytic interpretation of Means, B., Toyama, Y., Murphy, R., & Baki, M.(2013), which cautions against attributing blended-learning differences solely to the delivery medium.

The distinction between access, activity, performance, and learning is essential. Access means that educational opportunities are available. Activity refers to observable participation, including logins, messages, and task attempts. Performance means success on a particular observed task. Learning is inferred from changes in knowledge or skill that support subsequent performance. A system may facilitate access while delivering little durable learning; a learner may acquire understanding during quiet, unlogged reading; and high performance with immediate assistance may coexist with weak independent transfer.

Historical Formation of Technology-Mediated Education

The present landscape grew from several partially independent histories: organized distance education, educational broadcasting, programmed instruction, computer-assisted instruction, networked communication, and digital multimedia. Earlier distance formats established the organizational problem of supporting learners separated from a teacher. Computer-assisted approaches introduced adaptive sequences, immediate response evaluation, and individualized pacing. Networked environments later made communication among dispersed participants an ordinary feature of educational provision.

Early comparative work asked whether distance instruction could match conventional classroom instruction. In a major synthesis of research conducted between 1985 and 2002, Bernard, R. M., Abrami, P. C., Lou, Y., Borokhovski, E., Wade, A., Woz(2004) considered 232 studies and 688 outcomes covering achievement, attitudes, and retention. Their broadly near-zero aggregate differences coexisted with substantial heterogeneity. Historically, that combination mattered more than the apparent contest between delivery formats: educational arrangements varied enough for both successes and failures to coexist under the same general label.

As computer conferencing developed, researchers examined whether dispersed communication could support reasoned inquiry, mutual recognition, and instructional guidance. Garrison, D. R., Anderson, T., & Archer, W.(2000) articulated the Community of Inquiry framework in work grounded in text-based computer conferencing. It connected cognitive presence, social presence, and teaching presence to the quality of educational experience. Its explanatory contribution was to make the structure of interaction an object of analysis rather than treating online communication as an undifferentiated conduit.

Expansion of broadband access, learning-management systems, mobile computing, recorded video, and digital assessment widened the repertoire of possible instructional combinations. This expansion also multiplied confounders. The meta-analysis by Means, B., Toyama, Y., Murphy, R., & Baki, M.(2013) considered online and blended designs in a period when the latter often included additional instructional materials and learning time. Digital education became an empirical field concerned increasingly with interactions between technology, pedagogy, and institutional implementation.

The pandemic revealed a historically distinctive form of emergency remote teaching. Hodges, C., Moore, S., Lockee, B., Trust, T., & Bond, A.(2020) explained why rapid crisis substitution should be distinguished analytically from courses developed with sustained online instructional planning. This distinction protects the interpretability of evidence: outcomes during school closures reflect family circumstances, societal disruption, health threats, teacher preparation, assessment changes, and access constraints as well as delivery modality.

Theories of Learning in Mediated Environments

Cognitive processing and instructional organization

Cognitive accounts identify limitations in attention and working memory, the acquisition of organized knowledge, and the effects of practice. In Sweller, J.(1988) account of cognitive load during problem solving, learner resources devoted to means–ends search can compete with processes needed for schema acquisition. Digital lessons expose these issues through split attention, competing windows, redundant narration, uncontrolled pacing, and fragmented representations. Yet digital systems can also make examples, revisiting, pausing, and stepwise practice easier to organize.

The cognitive theory of multimedia learning proposed by Mayer, R. E., & Moreno, R.(2003) emphasizes verbal and pictorial processing, limited cognitive capacities, and active integration. The scientific implication is a design question: when does a combination of text, voice, animation, and interaction help a learner select, organize, and connect important information, and when does it overburden attention? Visual novelty is not a proxy for conceptual clarity.

Self-regulated learning

Self-regulated learning concerns how learners set goals, choose strategies, monitor understanding, allocate effort, and revise behavior. Zimmerman, B. J.(2002) describes goal setting, self-observation, strategic action, and self-evaluation as interacting processes. In online learning, the temporal and spatial separation of instruction often shifts decisions about scheduling, help seeking, and sustained attention toward the learner. This creates autonomy alongside new demands on planning and metacognition.

In a systematic review of online higher education, Broadbent, J., & Poon, W. L.(2015) found that time management, metacognition, effort regulation, and critical thinking were associated with achievement, while associations for particular strategies differed. Such correlational patterns identify plausible targets for instructional design; they do not demonstrate that selecting high-performing self-regulators causes their achievement or that a single intervention works equally across populations.

Social and relational accounts

Socially organized learning relies on explanation, dialogue, feedback, modeling, negotiation of meaning, and norms of participation. The Garrison, D. R., Anderson, T., & Archer, W.(2000) provides one influential language for studying these processes in technology-mediated settings. Social presence refers to the experience of others as participants in an interaction, teaching presence to the design and facilitation of instruction, and cognitive presence to purposeful construction and examination of meaning.

The model is useful for research because each dimension can guide measurement, but a course may score highly on perceived presence without proving gains on an external delayed test. This distinction anticipates a recurring theme: psychological experience and durable learning are associated educational outcomes requiring separate measurement.

Educational ecology and situated participation

Learning is also shaped by material resources, competing obligations, household conditions, organizational policies, and cultural expectations. The same asynchronous format may support a working adult who needs flexible time and burden a child without a quiet study space. A hybrid course can provide vital human contact while requiring transportation and predictable scheduling. An ecological interpretation therefore asks which learners can actually use the opportunities a course affords.

A comprehensive scientific explanation joins cognitive, interpersonal, institutional, and technical levels. Reducing learning success to either an individual's traits or a platform feature loses interactions that are central to educational experience.

Research Designs and Evidence Quality

A course-comparison study must specify the causal contrast. One possible comparison holds subject matter, instructional time, teacher support, and assessment constant while changing delivery format. Another replaces a traditional class with an integrated online program that also changes contact hours and feedback. These studies answer different questions. Random assignment improves causal interpretation within the population and implementation tested; observational comparisons require attention to prior achievement, choice of format, instructor assignment, socioeconomic access, and institutional conditions.

Outcome measures must be classified before they are combined. Course grades aggregate assignments, evaluation policies, persistence, and instructional assistance. A multiple-choice posttest measures performance on sampled items at a particular time. A delayed test examines retention; a structurally different task tests transfer; a portfolio can evaluate complex production while introducing rating reliability questions. Student satisfaction, belonging, engagement, self-reported learning, and log-based participation are additional dimensions. They are informative without being interchangeable.

Meta-analytic comparisons offer a map of average findings, but their validity depends on inclusion decisions, dependence among multiple outcomes from the same course, variability in interventions, risk of bias, missing negative studies, and the treatment of observational versus experimental designs. Bernard, R. M., Abrami, P. C., Lou, Y., Borokhovski, E., Wade, A., Woz(2004) explicitly reported substantial variability across applications. The U.S. Department of Education, Office of Planning, Evaluation, and Poli(2010) screened a large body of literature to identify a much smaller subset with usable comparative designs. This illustrates the difference between the volume of educational-technology publication and the amount of evidence suited for a narrowly defined causal estimate.

A further concern is interference: students assigned to different conditions may share notes, instructors may modify behavior across arms, and digital materials may be used outside the scheduled treatment. Subgroup analyses face small samples and selective reporting. Evidence from a single institution or introductory course may not generalize to children, learners with disabilities, laboratory training, languages, or resource-constrained settings. An adequate scientific synthesis therefore reports what was compared, who participated, what was measured, and the conditions under which findings could plausibly travel.

Comparative Effectiveness and the Medium Question

The question “Does online learning work?” can be answered only after its comparator and outcome are made explicit. Bernard, R. M., Abrami, P. C., Lou, Y., Borokhovski, E., Wade, A., Woz(2004) reported essentially zero overall mean differences across their broad historical literature, alongside considerable heterogeneity. Their result supports the proposition that delivery format does not determine achievement by itself. It offers no guarantee that a particular online program equals a particular classroom program.

The later Means, B., Toyama, Y., Murphy, R., & Baki, M.(2013) synthesis found modestly better average achievement in online conditions in the included comparisons, with a statistically significant advantage in blended-versus-face-to-face contrasts but not in wholly online-versus-face-to-face contrasts. Their analysis also notes that blended conditions frequently supplied extra instructional time, resources, or interactions. Thus an apparent advantage of a blend may reflect a larger or more carefully designed educational treatment. This is a confounding explanation supported by the study's treatment coding, not a reason to dismiss the observed outcomes.

A controlled illustration comes from Bowen, W. G., Chingos, M. M., Lack, K. A., & Nygren, T. I.(2014), who studied a hybrid statistics course across six public university campuses using random assignment. The study's design makes its estimates more informative than simple comparisons of students who voluntarily select different formats. It also establishes a narrow scope: an interactive machine-guided statistics course with a specified face-to-face component, not all online education. Course-specific outcomes and efficiency must be considered in light of the instructional resources and particular student population.

The cumulative finding is that educational interventions can be effective in different media. Estimates obtained from historical correspondence courses, interactive online statistics, pandemic emergency remote schooling, and contemporary multimodal systems represent different treatment families. There is no psychologically plausible universal medium coefficient that remains invariant across subjects, learners, assessments, and institutions.

Synchronous and Asynchronous Organization

Synchronous instruction supports immediate questioning, spontaneous discussion, shared attention, and time-bounded group activity. It can also impose scheduling burdens, connectivity requirements, fatigue from continuous videoconferencing, and limited opportunities for slow reflection. Asynchronous instruction enables revisiting, deliberation, flexible pacing, and participation by learners managing work or caregiving. It increases the importance of explicit task structure, response expectations, time management, and delayed feedback pathways.

Hrastinski, S.(2008) argued that synchronous and asynchronous activities support somewhat different purposes, including community development and reflective participation. These are functional distinctions that can guide course design. A real-time meeting is not inherently dialogic, and a discussion forum is not inherently reflective. Both require tasks that elicit meaningful educational activity.

The meta-analysis by Bernard, R. M., Abrami, P. C., Borokhovski, E., Wade, C. A., Tamim, R.(2009) examined interaction treatments in distance education and found gains associated with stronger designed opportunities for student–content, student–teacher, or student–student interaction. Its results are especially relevant to asynchronous contexts, where communication and engagement must often be made explicit through instructional structure. The authors reported substantial heterogeneity, which prevents the pooled outcome from being interpreted as the effect of a single universal instructional technique.

A sound allocation of activities asks how learning goals relate to time. Concept review and low-stakes retrieval can benefit from asynchronous spacing. Collaborative negotiation, diagnostic clarification, oral practice, and responsive coaching may benefit from synchronous exchange. Many courses need both, including carefully designed routes for learners who cannot attend at a particular hour.

Blended Architecture and Educational Integration

Blended learning is scientifically significant when its components are joined through purposeful instructional sequencing. An online preparation activity may introduce foundational concepts, a classroom session may develop explanation and application, and subsequent digital practice may support retrieval and feedback. The critical variable is how the combined arrangement creates opportunities that neither component alone would necessarily provide.

Blending can generate extra work for teachers and students as well as useful flexibility. Duplicated deadlines, redundant videos, fragmented communication channels, and inconsistent assessment policies create costs that may be hidden in a broad classification such as “hybrid.” Comparing blended with classroom instruction requires recording total contact time, independent study demands, instructional resources, and the distribution of feedback. Otherwise claims of greater effectiveness conflate an enriched program with a change of delivery.

The favorable blend comparisons discussed by Means, B., Toyama, Y., Murphy, R., & Baki, M.(2013) are best interpreted as evidence that some integrated instructional arrangements can improve outcomes, with unresolved attribution to their particular ingredients. A useful research design factorially manipulates elements such as instructional medium and amount of guided practice, or tests alternative blends with equalized time and resources.

Blended design also has a psychological continuity problem. Learners must perceive how online preparation, live discussion, group work, and assessment belong to the same conceptual sequence. When the parts are disconnected, participation becomes compliance with parallel tasks. When the parts are aligned, learners can use knowledge from one context to act in another, creating a stronger basis for transfer tests.

Interaction, Teaching Presence, and Social Presence

Three forms of designed interaction

Interaction is a central explanatory variable in mediated education. Student–content interaction occurs when learners actively interpret, manipulate, practice, and test material. Student–teacher interaction supplies instructional framing, correction, assessment, and emotional and intellectual guidance. Student–student interaction includes collaborative explanation, questioning, debate, and the coordination of shared tasks. These categories organize interventions, but their effects depend on quality, timing, and alignment with the educational objective.

In their meta-analysis of 74 achievement contrasts, Bernard, R. M., Abrami, P. C., Borokhovski, E., Wade, C. A., Tamim, R.(2009) reported a methodological-quality-adjusted weighted average effect of 0.38 for interaction treatments, with heterogeneity. This estimate concerns the studies and intervention contrasts included, rather than a guaranteed improvement from installing a forum or adding teacher messages. Strong interactions are designed and enacted educational processes, whereas mere opportunities to click or communicate provide a weak operational definition.

The empirical problem is to distinguish nominal from consequential interaction. An instructor may reply frequently without identifying misconceptions. Learners may post many agreeable messages without evaluating claims. A quiz may display correctness while leaving the source of the error unexplained. Studies should therefore evaluate interaction content, discourse quality, feedback specificity, and subsequent independent performance, as well as counting events.

Community of Inquiry and its measurement

The framework introduced by Garrison, D. R., Anderson, T., & Archer, W.(2000) presents cognitive, social, and teaching presence as interdependent features of online educational experiences. The original work used indicators from computer-conferencing transcripts. Later Arbaugh, J. B., Cleveland-Innes, M., Diaz, S. R., Garrison, D. R., Ice(2008) contributed survey development and validation, enabling broader quantitative investigation. These measures help researchers describe educational perceptions and patterns of participation.

However, construct validity requires a clear distinction between the dimensions measured and their hypothesized effects. Teaching presence measured as student perception may reflect responsiveness, clarity, and assessment structure, while actual instructor actions require behavioral observation or course analysis. Cognitive presence ratings are related to the perception of inquiry and understanding, whereas delayed mastery tests directly sample later knowledge. Triangulation improves interpretation.

Richardson, J. C., Maeda, Y., Lv, J., & Caskurlu, S.(2017) synthesized evidence linking social presence to satisfaction and perceived learning. Such results establish a consequential relationship within self-report research. Shared method variance, the direction of causation, and the difference between perceived and demonstrated learning remain important. Students who are more successful may feel more connected; courses with better teaching may produce both connection and achievement.

Learner participation and social agency

Students also contribute to the environment's organization. The Shea, P., & Bidjerano, T.(2010) formulation of learning presence links self-regulation and self-efficacy to online inquiry. Their study examined ratings among 3,165 students from 42 institutions. Its broad institutional sample makes the analysis informative, but ratings gathered within the same survey framework cannot by themselves establish the causal contribution of any learner attribute.

A productive digital community is a structured field of reciprocal actions: participants receive a clear purpose, have room to ask questions, get timely and intelligible responses, and can contribute without fear of humiliation. This conception applies to classrooms and online spaces, with technology changing the visibility, timing, persistence, and accessibility of participation.

Cognitive Load and Multimedia Design

Multimedia instruction can distribute explanatory content across narration, text, diagrams, worked examples, animation, and learner controls. Its design depends on cognitive processing constraints. Mayer, R. E., & Moreno, R.(2003) articulated a theory based on the selection and integration of verbal and pictorial information under limited capacity. The scientific point is that additional channels are helpful when they support construction of a coherent representation, and burdensome when they create competing processing demands.

A lecture video can be segmented into meaningful units with opportunities for retrieval and reflection. Captions enable access and can support comprehension, yet overlapping text, voice, and animation may require careful timing. Visual emphasis can direct attention to relevant relations; decorative motion can compete for the same capacity. The learner's prior knowledge matters because an explanation that helps a novice may become redundant for an expert.

In their synthesis of randomized trials, Noetel, M., Griffith, S., Delaney, O., Sanders, T., Parker, P., del Po(2021) reviewed 105 eligible studies involving 7,776 students in higher education. They reported different pooled effects when video replaced an existing method versus when it was added, demonstrating that the treatment contrast matters. The abstract also acknowledges possible experimental and publication biases. The evidence supports the potential usefulness of educational video without establishing that longer videos, greater production expense, or audiovisual saturation necessarily improve learning.

Design experiments should isolate relevant ingredients: signaling versus no signaling, guided segmentation versus continuous presentation, explanatory diagrams versus decorative images, video plus retrieval versus video alone. The learner's ability to pause or replay should be measured as an affordance and an observed behavior, with its association to outcomes tested cautiously. Long-term retention requires an outcome measured after a meaningful interval.

The medium itself may change the perceived effort of a task. Fluency during a polished presentation can feel like mastery, while active problem solving produces difficulty. Deslauriers, L., McCarty, L. S., Miller, K., Callaghan, K., & Kestin, (2019) showed in classroom experiments that actual learning and students' feeling of learning can diverge. The result is relevant to digital learning because satisfaction ratings can favor smooth presentations even when effortful activity yields better tested understanding.

Self-Regulation, Metacognition, and Learner Autonomy

The temporal flexibility of digital learning changes who controls the schedule. A learner may be able to revisit material, reorganize practice, and choose when to request support. This flexibility is valuable when study strategies are effective and problematic when work is delayed, distractions accumulate, or feedback is too distant. The relevant psychological construct is regulation of learning activity within constraints.

Zimmerman, B. J.(2002) account emphasizes planning, self-monitoring, and reflection. A digital course can operationalize these processes through clear goals, manageable modules, formative checkpoints, calendars, reflection prompts, and opportunities to diagnose errors. The goal is to support independent educational agency rather than maximize dependence on reminders and automated prompts.

The systematic review by Broadbent, J., & Poon, W. L.(2015) identifies correlations between achievement and time management, metacognition, effort regulation, and critical thinking in online higher education. Interpretation requires caution because stronger students can differ across multiple unmeasured characteristics, and course design itself can shape both strategy use and grades. Interventions that assign learners to structured planning, spacing, or reflection conditions provide stronger evidence about causal effects than comparisons of survey scores.

Self-monitoring is also vulnerable to illusions of competence. Learners can mistake recognition for retrieval, repeated exposure for acquisition, and quick completion for durable understanding. In their comprehensive evaluation of study techniques, Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham(2013) emphasized the utility of practice testing and distributed practice across various educational settings. Digital tools make these techniques scalable, but their implementation must avoid trivial questions, immediate cue dependence, and practice tasks that reproduce the final test too narrowly.

A strong autonomy-supporting system gradually transfers organizational responsibility toward the learner. Early scaffolds can teach how to plan and check understanding; later stages can require students to select strategies, explain decisions, and perform without prompts. Independence is an empirical outcome measured through unassisted learning and transfer, rather than a rhetorical property of technology.

Motivation, Engagement, and Belonging

Engagement describes behavior, cognitive investment, and affective involvement in learning. These dimensions can align but also diverge. Students can spend extensive time online while confused or distracted, contribute few posts while studying carefully, or report strong satisfaction without retaining the material. A scientific evaluation needs independent outcome measures rather than treating a login count as the definitive marker of motivation.

Online environments can support motivational needs through meaningful choices, visible progress, successful mastery experiences, reciprocal feedback, and social connection. They can also intensify uncertainty when learners encounter opaque instructions, delayed responses, isolation, or poorly coordinated tasks. The causal mechanisms are task structure, perceived competence, anticipated value, social relationships, and opportunity, acting within a particular learner's circumstances.

The meta-analysis of Richardson, J. C., Maeda, Y., Lv, J., & Caskurlu, S.(2017) shows strong positive associations between social presence and satisfaction and perceived learning. The distinction from externally assessed learning must remain visible. A more recent meta-analysis by Doo, M. Y., & Kim, J.(2024) examined the engagement–outcome relation in online higher education across 34 eligible studies. Its synthesis contributes evidence about associations and moderators, but the association alone does not settle whether engagement produces achievement or reflects other strengths of students and courses.

Belonging also depends on how digital classrooms distribute voice. Persistent written discussion can give a reflective learner more time to formulate a contribution. Live audiovisual participation can strengthen immediacy for some students while creating anxiety or excluding those with unstable connectivity. Instructional design should provide multiple meaningful participation routes and make expectations transparent.

An important practical implication is that metrics should serve diagnostic questions. A sudden decline in assignment submission may call for communication about workload or access. A high rate of video completion coupled with weak transfer may call for deeper practice tasks. A student's emotional connection to the course deserves attention even when its measurement cannot substitute for a knowledge assessment.

Assessment, Feedback, and Retrieval

Assessment in mediated settings performs several functions: estimating present knowledge, diagnosing misconceptions, guiding study, supporting certification, and documenting progress. These functions have different validity requirements. An automatically graded factual item can support low-stakes practice, whereas complex reasoning, professional judgment, and collaborative products may require richer evaluation. Assessment design must consider whether the intended construct is actually measured.

Hattie, J., & Timperley, H.(2007) explain that feedback effects depend on their content and use. Information about errors, goals, strategies, and next steps can support learning; feedback reduced to generic praise or constant correctness signals may provide little guidance. Online courses can deliver rapid responses at scale, but quality depends on the nature of the response and the opportunity for learners to act on it.

A review by Gikandi, J. W., Morrow, D., & Davis, N. E.(2011) emphasized that online formative assessment can promote learner-centered activity and meaningful feedback while raising questions about validity, reliability, and authenticity. Opportunities include quizzes, peer responses, discussion, and portfolios. The strongest interpretation is that these are possible mechanisms for improvement, provided the educational sequence elicits substantive thinking.

Retrieval practice creates a special case because an assessment event can also be a learning event. Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham(2013) identify practice testing and distributed practice as broadly useful learning techniques. In digitally delivered courses, spaced retrieval is easy to schedule and personalize, yet effectiveness depends on question quality, effortful recall, corrective feedback, spacing intervals, and whether learners subsequently demonstrate independent understanding.

Academic integrity and measurement are closely related. When assignments can be completed with extensive outside assistance, a score may reflect the combined human–tool system rather than the learner's own knowledge. Authentic assignments, explanation of reasoning, oral defenses, controlled practice, and delayed unassisted assessments can each contribute complementary evidence. None is a universally sufficient detector of unauthorized help; assessment policy should make legitimate forms of assistance and attribution explicit.

Attention, Distraction, and Digital Multitasking

Digital interfaces contain instructional and noninstructional affordances simultaneously. A learner may encounter lecture material, messages, notifications, search results, unrelated media, and automated suggestions on one device. Task switching requires attentional reorientation and may fragment the processing needed to build coherent knowledge. The psychological issue is the allocation and control of attention, not a categorical judgment about screens.

Sana, F., Weston, T., & Cepeda, N. J.(2013) experimentally investigated laptop multitasking in a simulated lecture. They found poorer test performance among multitasking students and among nearby students able to view a multitasking peer. The controlled design supports a causal interpretation for that particular interference, setting, and task. Its findings should not be generalized automatically to every use of laptops, digital note taking, assistive technology, or independent online study.

Distraction can arise from interface design, social expectations, motivational state, workload, and environmental conditions. Institutional policies requiring uninterrupted camera presence may themselves introduce competing demands. A scientifically defensible approach measures the source, frequency, and timing of task switching and its relation to actual learning rather than presuming that all concurrent device activity is equivalent.

The response is instructional as well as behavioral. Shorter coherent learning units, clear task goals, voluntary notification controls, scheduled pauses, accessible materials, and purposeful interaction can improve the opportunity for sustained attention. Researchers should distinguish the effects of such design elements from individual differences in prior knowledge, attention regulation, and competing obligations.

Access, Disability, and Educational Equity

Access is a multidimensional condition. It includes hardware and network availability, bandwidth and data costs, usable learning spaces, reliable electricity, time, accessible interfaces, appropriate language, instructional support, and the capacity to respond to inevitable disruptions. Counting devices or accounts alone underestimates barriers experienced by people with disabilities, learners in rural locations, students with caregiving responsibilities, and those whose institutional environments provide limited support.

The UNESCO.(2023) documents uneven deployment, variable evidence of educational effects, and the importance of equitable, evidence-based, sustainable use. As a policy synthesis, it provides an international perspective; it is not an experiment demonstrating one globally uniform treatment effect. Its value is in showing that digital education must be evaluated within material and institutional systems.

Accessible learning should provide meaningful alternatives for receiving information, participating, and demonstrating knowledge. Captions, transcripts, screen-reader support, keyboard navigation, readable contrast, flexible timing, and accessible documents address different needs. The presence of a technical feature does not prove that the learning activity is accessible: a caption may be inaccurate, an interactive diagram may lack a usable textual equivalent, and a timed quiz may impose avoidable barriers.

Equity is not identical to uniformity of delivery. Providing every student with the same live video schedule can disadvantage learners with unstable connectivity or conflicting obligations. Asynchronous options may improve flexibility while increasing demands on literacy and self-regulation. A good design evaluates equivalent opportunities for learning, communication, practice, and assessment and checks whether different groups can use them.

Causal equity research requires both subgroup outcomes and measures of differential access to the intervention. Selective dropout can produce an apparently successful course when those facing the greatest barriers disappear from the analytic sample. Meaningful policy analysis tracks enrollment, participation, completion, delayed achievement, and learners' reported barriers, with appropriate privacy safeguards.

Developmental and Institutional Variation

Developmental stage changes the resources available for self-regulation, abstract reasoning, reading, help seeking, and independent scheduling. Children may require more explicit routines, caregiver coordination, and teacher supervision. Adolescents face changing peer relationships, motivational priorities, and competing digital activities. Adults may benefit from scheduling flexibility while combining study with employment and family responsibilities. These are broad design considerations whose application should be examined empirically in each population.

Educational objectives also differ. Foundational reading and numeracy, laboratory practice, clinical skill training, collaborative design, language conversation, and conceptual university coursework require different evidence of success. A program effective for structured factual practice cannot be assumed to provide all conditions for supervised professional performance. Claims of general effectiveness should therefore identify the knowledge domain, developmental level, and form of assessment.

Institutions shape outcomes through instructor preparation, support services, assessment rules, platform reliability, program coordination, and the allocation of educational time. The review by Martin, F., Sun, T., & Westine, C. D.(2020) found a concentration of research on learners and engagement and called attention to less developed organizational-level study. This imbalance matters because decisions about instructional provision and learner support are frequently made at the institutional level.

The quality of implementation is not incidental. A course may use research-supported techniques on paper yet provide sparse feedback, inaccessible content, or little instructor time. Implementation fidelity, adaptation, and local context should be measured rather than invoked only after outcomes diverge from expectations.

Emergency Remote Education and Educational Disruption

Emergency remote education is an instructional response to disruption under limited planning time. It can involve online teaching, distribution of printed materials, radio or television, temporary schedules, and improvised assessment. The critical distinction introduced by Hodges, C., Moore, S., Lockee, B., Trust, T., & Bond, A.(2020) is methodological: outcomes observed during crisis substitution cannot automatically be generalized to carefully planned online programs.

School closures during the COVID-19 pandemic constitute an important research setting because the educational change occurred alongside exceptional social conditions. Learning opportunities were affected by institutional closures, health concerns, household resources, parental support, unequal connectivity, shifting curricula, and changing assessment practices. A study that compares cohorts before and during the pandemic estimates a composite historical disruption unless its design credibly separates component pathways.

Betthäuser, B. A., Bach-Mortensen, A. M., & Engzell, P.(2023) conducted a preregistered systematic review and meta-analysis of learning during the pandemic using 42 studies across 15 countries. Their synthesis documents substantial concern about learning deficits and uneven effects, while naturally inheriting the identification limits and variation of the included study designs. The correct interpretation is that educational disruption was associated with measurable losses in many studied settings, not that planned online instruction has a single intrinsically negative effect.

The scientific unit of analysis should include both instruction and the conditions that enable it. Disruptions may alter time on task, contact with teachers, opportunities to practice, psychological safety, and access to testing. The observed outcome depends on the accumulation of these changes. Recovery research can compare targeted tutoring, additional instructional time, formative diagnosis, and restored social support, while avoiding a technology-centered explanation where multiple causes operate.

Emergency preparedness also belongs to educational design. Institutions can maintain accessible materials, alternative communication channels, transparent assessment policies, and support for interrupted study. Such provisions have value for pandemics, displacement, weather emergencies, disability-related absence, and other interruptions without assuming that any one format suits every situation.

Learning Analytics and Measurement Validity

Digital environments create records of clicks, timestamps, attempts, responses, messages, and content access. These records can support formative assessment, institutional monitoring, and research on learning processes. They also create a danger of mistaking trace availability for construct validity. The best-observed behaviors are those the platform happens to record, not necessarily the behaviors most important to understanding.

A log entry may show that a video player was open, but it cannot establish attention to the explanation. Time spent on a problem may reflect thoughtful struggle, confusion, distraction, or interruption. Frequent forum posts may indicate collaboration, assessment compliance, or social performance. A learner's recorded absence may reflect connectivity failure rather than low motivation. The meaning of a trace must be validated against task context and independent outcomes.

Educational measurement benefits from a construct map specifying observable indicators, plausible alternative explanations, and external evidence. If a platform predicts course grades, investigators must ask whether both variables are driven by the same assignment rules. If a prediction model is trained on students from one institution and evaluated on the same institutional calendar, its performance may be inflated by patterns that fail in a different setting. Predictive accuracy is also distinct from the effects of interventions selected using those predictions.

Longitudinal designs can track changes in engagement, achievement, and use of assistance. Multilevel models may account for students nested in courses and institutions. Experiments may compare feedback policies or adaptive support. Qualitative interviews and observations can explain why the same recorded activity occurs under different learning conditions. Combining these methods can produce a more faithful account of psychological mechanisms than telemetry alone.

Ethical measurement requires a purpose proportionate to the data collected, transparent interpretation, access controls, retention limits, and attention to the effects of labels placed on students. Students should not be characterized by a single predictive score as though it were a fixed psychological property. Error rates and differential performance across groups require analysis before analytics are used to allocate opportunities or penalties.

Artificial Tutoring and Human–Artificial Learning

Artificial tutoring changes the availability, timing, and form of educational assistance. Conventional tutoring systems can sequence problems, provide rule-based hints, and adapt practice; generative language systems can also formulate explanations, simulate questioning, summarize materials, and respond in natural language. The educational object remains the human learner's acquisition, independent application, and durable retention of knowledge. The behavioral capabilities of an artificial tutor are examined here as components of the environment in which learning occurs.

Kasneci, E., Sessler, K., Küchemann, S., Bannert, M., Dementieva, D., (2023) discuss opportunities and risks associated with large language models in education, including personalized content and interaction together with factual unreliability, bias, overreliance, and the need for critical evaluation. Their paper is a commentary and conceptual assessment, not a randomized demonstration of universal effectiveness. This distinction is essential when new educational technology is described as transformative before its learning effects have been independently established.

Direct experimental evidence is developing. In a randomized study with 194 students in an undergraduate physics course, Kestin, G., Miller, K., Klales, A., Milbourne, T., & Ponti, G.(2025) compared a custom pedagogically designed artificial tutor with active-learning classroom instruction. The authors reported stronger learning outcomes in less time for their tutor condition and favorable learner perceptions. The result is important as an intervention-specific causal comparison. It should not be interpreted as the effect of all language-model tutors, all subjects, all age groups, or long-term independent retention without further testing.

Assistance, agency, and the problem of unassisted transfer

A learner who completes a task with continuous artificial assistance may show excellent immediate performance while developing less independent competence. The key empirical test is whether performance survives reduction of assistance, changed problem structure, delayed measurement, and unfamiliar contexts. Researchers can randomize levels of hint availability, answer disclosure, feedback timing, or tutoring intensity; measure unassisted performance afterward; and examine whether benefits persist.

The design of assistance matters. A tutor may ask diagnostic questions and prompt retrieval, supply partial hints, provide complete solutions, or simply generate finished work. These treatments should be studied separately. Guidance that elicits the student's own explanation has a different hypothesized mechanism from answer replacement. The same system may also help teachers create examples and identify misconceptions, without proving that its generated materials are accurate or educationally suitable.

Human–artificial learning introduces issues of epistemic trust. A fluent explanation can appear authoritative despite factual error. Verification should therefore be integrated into instruction: students can compare sources, challenge assumptions, test a solution against data, and document when assistance was used. The practical objective is an educational environment that expands the learner's competence and judgment rather than reducing assessment to assisted completion.

Artificial Era as a research horizon

Artificial Era · From Homo to Artificial identifies a wider historical context in which artificial systems become durable participants in educational and cultural arrangements. In this monograph, the category is an authorial intellectual framing, while empirical claims are restricted to observed instructional behavior, educational outcomes, and independently testable models. Claims about the psychological organization or subjective experience of artificial systems require their own definitions and evidence.

The research agenda includes distributed cognition between humans and artificial assistants, pedagogical dependence, calibration of trust, continuity of educational relationships, and changes in the distribution of explanatory work. These problems connect educational psychology with Psychology of Human–Artificial Relations and Psychology of Artificial. Their relation to this dossier lies in the measurable effects of mediated instruction on human learning.

Causal Explanations, Contradictions, and Limitations

The literature contains several results that can appear inconsistent until the unit of comparison is examined. Early distance-education meta-analysis reported essentially zero average modality differences with substantial dispersion; later online-learning syntheses reported modest overall advantages, especially in blended conditions; pandemic studies documented learning setbacks; controlled artificial tutoring studies can show advantages in carefully specified courses. These findings address different populations, treatments, outcomes, and historical conditions. A single verdict on “online versus classroom” would suppress the differences that give each study meaning.

One causal difficulty is treatment bundling. A blended program may introduce digital access, additional study time, new instructional materials, automated feedback, and redesigned teaching simultaneously. If learning improves, the package may be effective while the responsible ingredient remains uncertain. Component experiments, dismantling studies, and equal-time comparisons would help identify mechanisms. Measures of actual instructor effort and student workload should accompany achievement outcomes.

Another difficulty is the sampling of educational outcomes. An immediate multiple-choice examination can reward short-lived familiarity, while a delayed transfer assessment may produce a different ranking. Deslauriers, L., McCarty, L. S., Miller, K., Callaghan, K., & Kestin, (2019) demonstrate that perceived and measured learning can diverge even in tightly studied classroom settings; digital-course satisfaction and platform engagement thus require independent validation. Educational evaluations should register which endpoint answers their research question before interpreting success.

Selection bias is common in observational comparisons. Students choose modes according to employment, commuting constraints, prior achievement, confidence, access, and family responsibilities. Instructors and institutions select which courses to offer digitally. Propensity adjustment can reduce measured confounding but cannot eliminate unmeasured causes. Randomization is informative but often constrained by voluntary participation, instructor variation, intervention contamination, and limited external validity.

Publication bias and temporal change further complicate generalization. Technological systems, student expectations, institutional capacity, and assessment practices evolve. Results of a particular platform version or historical educational emergency may not transport to contemporary teaching contexts. Replication across institutions, languages, socioeconomic settings, and age groups is therefore as important as increasing the sample size at one campus.

Psychometric issues shape conclusions about engagement and presence. Shared self-report methods can inflate associations among satisfaction, motivation, and perceived learning. Log-derived activity metrics can misclassify invisible study. Course grades may incorporate attendance, participation, and completion rules differing across treatment arms. Researchers should identify the construct represented by each measure and test for invariance across populations and platforms.

Equity may also modify average findings. If certain learners cannot reliably attend video sessions, the effect of synchronous instruction is conditional on access. If a program loses participants who experience substantial barriers, reported success among completers can exaggerate its usefulness for the enrolled population. Intention-to-treat analysis, attrition reporting, subgroup analysis with appropriate power, and sensitivity analysis strengthen the evidential base.

A theoretical synthesis emerges from these considerations. The effective educational unit is a configuration of tasks, people, media, resources, and temporal constraints. Psychological mechanisms such as attention, retrieval, explanation, monitoring, motivation, and social interaction operate through that configuration. Research advances when each mechanism is operationalized and its contribution is tested independently or through well-defined interactions.

Research Agenda and Design Implications

The next stage of research should replace broad labels with explicit instructional components. Reporting ought to describe contact time, time spent on independent work, instructor support, feedback modality, assessment design, synchronous and asynchronous participation requirements, accessibility provisions, and the degree of artificial assistance. This would make replications and meta-analytic comparisons more interpretable.

Longitudinal outcome designs are particularly valuable. They should measure retention after meaningful delays, transfer to altered tasks, independent problem solving after assistance is withdrawn, and learner confidence calibrated against actual performance. In educational systems where short-term metrics are easy to obtain, these endpoints require deliberate attention and appropriate support for participants.

Better causal work should test interactions among learner characteristics, institutional support, and educational design. A strategy effective for highly self-regulated university students may require additional scaffolding for novices. Such moderation analyses need prespecified hypotheses, suitable statistical power, credible measurement, and replication. Their purpose is to identify functional educational conditions, not to produce permanent labels for individuals.

Research on online community and presence should pair perceptions with observed discourse, instructional behavior, and external achievement. In particular, interventions that improve social connection can be examined for their effects on participation, persistence, conceptual change, and delayed learning separately. These outcomes form a causal sequence that empirical research should test rather than assume.

Educational access research should compare the experience of those who enroll, those who remain, and those who are unable to participate. Accessibility audits, participatory design, infrastructure records, and educational outcomes can jointly identify barriers. Privacy-conscious analysis of digital traces should acknowledge missing data and avoid using convenience of measurement as a substitute for construct validity.

Artificial tutoring research should independently replicate experimental results, specify tutor architecture and instructional prompts, examine hallucination exposure, evaluate fairness and accessibility, and compare guided explanation with direct solution generation. The strongest long-term outcome is the learner's demonstrated ability to understand, evaluate, and act without the system, while retaining the freedom to use appropriate tools in authentic practice.

For instructional design, the current evidence supports several research-grounded priorities: begin from the knowledge to be learned; choose communication modes according to task demands; manage cognitive processing; provide purposeful interaction and informative feedback; make retrieval and review opportunities available; create routes for flexible and accessible participation; and assess learning independently of activity traces. These are hypotheses and evidence-based design principles whose concrete effects remain subject to local evaluation.

Conclusion

Digital, online, and blended learning environments are psychological and institutional arrangements through which instruction, interaction, practice, and assessment are organized. Their educational properties emerge from the coupling of instructional design with learner knowledge, self-regulation, social relationships, material access, and institutional support. A digital medium expands certain opportunities, changes the cost of others, and makes new traces of educational activity available. Each effect requires an explicit mechanism and an appropriate measure.

The strongest cumulative evidence rejects a universal evaluation of educational quality from format labels alone. Historical comparative meta-analyses demonstrate wide heterogeneity; research on interaction, multimedia, self-regulation, and formative assessment identifies plausible mechanisms; carefully controlled course trials show what specified interventions can accomplish; pandemic evidence demonstrates the importance of historical and material context; and current artificial tutoring experiments open a new class of testable instructional conditions. The interpretation of these bodies of evidence depends on separating observational associations, randomized effects, learner perceptions, and independently demonstrated knowledge.

A scientifically mature account of digitally mediated education will measure durable knowledge, transfer, access, agency, and participation while documenting the instructional conditions that produced them. This account also establishes the appropriate basis for studying education in the Artificial Era: the transformation of educational relationships and resources is examined through verifiable outcomes, explicit constructs, and causal research designs. The resulting research program treats learning as a sustained human capacity whose development is shaped by the organization of educational environments.

References

  • Arbaugh, J. B., Cleveland-Innes, M., Diaz, S. R., Garrison, D. R., Ice, P., Richardson, J. C., Shea, P., & Swan, K. P. (2008). Community of Inquiry Framework: Validation and instrument development. The International Review of Research in Open and Distributed Learning, 9(2). https://doi.org/10.19173/irrodl.v9i2.573
  • Bernard, R. M., Abrami, P. C., Borokhovski, E., Wade, C. A., Tamim, R. M., Surkes, M. A., & Bethel, E. C. (2009). A meta-analysis of three types of interaction treatments in distance education. Review of Educational Research, 79(3), 1243–1289. https://doi.org/10.3102/0034654309333844
  • Bernard, R. M., Abrami, P. C., Lou, Y., Borokhovski, E., Wade, A., Wozney, L., Wallet, P. A., Fiset, M., & Huang, B. (2004). How does distance education compare with classroom instruction? A meta-analysis of the empirical literature. Review of Educational Research, 74(3), 379–439. https://doi.org/10.3102/00346543074003379
  • Betthäuser, B. A., Bach-Mortensen, A. M., & Engzell, P. (2023). A systematic review and meta-analysis of the evidence on learning during the COVID-19 pandemic. Nature Human Behaviour, 7, 375–385. https://doi.org/10.1038/s41562-022-01506-4
  • Bowen, W. G., Chingos, M. M., Lack, K. A., & Nygren, T. I. (2014). Interactive learning online at public universities: Evidence from a six-campus randomized trial. Journal of Policy Analysis and Management, 33(1), 94–111. https://doi.org/10.1002/pam.21728
  • Broadbent, J., & Poon, W. L. (2015). Self-regulated learning strategies & academic achievement in online higher education learning environments: A systematic review. The Internet and Higher Education, 27, 1–13. https://doi.org/10.1016/j.iheduc.2015.04.007
  • Deslauriers, L., McCarty, L. S., Miller, K., Callaghan, K., & Kestin, G. (2019). Measuring actual learning versus feeling of learning in response to being actively engaged in the classroom. Proceedings of the National Academy of Sciences, 116(39), 19251–19257. https://doi.org/10.1073/pnas.1821936116
  • Doo, M. Y., & Kim, J. (2024). The relationship between learning engagement and learning outcomes in online learning in higher education: A meta-analysis study. Distance Education, 45(1), 60–82. https://doi.org/10.1080/01587919.2024.2303484
  • Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students’ learning with effective learning techniques: Promising directions from cognitive and educational psychology. Psychological Science in the Public Interest, 14(1), 4–58. https://doi.org/10.1177/1529100612453266
  • Garrison, D. R., Anderson, T., & Archer, W. (2000). Critical inquiry in a text-based environment: Computer conferencing in higher education. The Internet and Higher Education, 2(2–3), 87–105. https://doi.org/10.1016/S1096-7516(00)00016-6
  • Gikandi, J. W., Morrow, D., & Davis, N. E. (2011). Online formative assessment in higher education: A review of the literature. Computers & Education, 57(4), 2333–2351. https://doi.org/10.1016/j.compedu.2011.06.004
  • Hattie, J., & Timperley, H. (2007). The power of feedback. Review of Educational Research, 77(1), 81–112. https://doi.org/10.3102/003465430298487
  • Hodges, C., Moore, S., Lockee, B., Trust, T., & Bond, A. (2020, March 27). The difference between emergency remote teaching and online learning. EDUCAUSE Review. https://er.educause.edu/articles/2020/3/the-difference-between-emergency-remote-teaching-and-online-learning
  • Hrastinski, S. (2008). Asynchronous and synchronous e-learning. EDUCAUSE Quarterly, 31(4), 51–55. https://er.educause.edu/articles/2008/11/asynchronous-and-synchronous-elearning
  • Kasneci, E., Sessler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser, U., Groh, G., Günnemann, S., Hüllermeier, E., Krusche, S., Kutyniok, G., Michaeli, T., Nerdel, C., Pfeffer, J., Poquet, O., Sailer, M., Schmidt, A., Seidel, T.,. Kasneci, G. (2023). ChatGPT for good? On opportunities and challenges of large language models for education. Learning and Individual Differences, 103, 102274. https://doi.org/10.1016/j.lindif.2023.102274
  • Kestin, G., Miller, K., Klales, A., Milbourne, T., & Ponti, G. (2025). AI tutoring outperforms in-class active learning: An RCT introducing a novel research-based design in an authentic educational setting. Scientific Reports, 15, 17458. https://doi.org/10.1038/s41598-025-97652-6
  • Martin, F., Sun, T., & Westine, C. D. (2020). A systematic review of research on online teaching and learning from 2009 to 2018. Computers & Education, 159, 104009. https://doi.org/10.1016/j.compedu.2020.104009
  • Mayer, R. E., & Moreno, R. (2003). Nine ways to reduce cognitive load in multimedia learning. Educational Psychologist, 38(1), 43–52. https://doi.org/10.1207/S15326985EP3801_6
  • Means, B., Toyama, Y., Murphy, R., & Baki, M. (2013). The effectiveness of online and blended learning: A meta-analysis of the empirical literature. Teachers College Record, 115(3), 1–47. https://doi.org/10.1177/016146811311500307
  • Noetel, M., Griffith, S., Delaney, O., Sanders, T., Parker, P., del Pozo Cruz, B., & Lonsdale, C. (2021). Video improves learning in higher education: A systematic review. Review of Educational Research, 91(2), 204–236. https://doi.org/10.3102/0034654321990713
  • Richardson, J. C., Maeda, Y., Lv, J., & Caskurlu, S. (2017). Social presence in relation to students’ satisfaction and learning in the online environment: A meta-analysis. Computers in Human Behavior, 71, 402–417. https://doi.org/10.1016/j.chb.2017.02.001
  • Sana, F., Weston, T., & Cepeda, N. J. (2013). Laptop multitasking hinders classroom learning for both users and nearby peers. Computers & Education, 62, 24–31. https://doi.org/10.1016/j.compedu.2012.10.003
  • Shea, P., & Bidjerano, T. (2010). Learning presence: Towards a theory of self-efficacy, self-regulation, and the development of a communities of inquiry in online and blended learning environments. Computers & Education, 55(4), 1721–1731. https://doi.org/10.1016/j.compedu.2010.07.017
  • Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12(2), 257–285. https://doi.org/10.1016/0364-0213(88)90023-7
  • U.S. Department of Education, Office of Planning, Evaluation, and Policy Development. (2010). Evaluation of evidence-based practices in online learning: A meta-analysis and review of online learning studies (revised September 2010). U.S. Department of Education. https://www.ed.gov/media/document/evaluation-of-evidence-based-practices-online-learning-meta-analysis-and-review-of-online-learning-studies-revised-september-2010-107159.pdf
  • UNESCO. (2023). Global education monitoring report 2023: Technology in education—A tool on whose terms? UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000385723
  • Zimmerman, B. J. (2002). Becoming a self-regulated learner: An overview. Theory Into Practice, 41(2), 64–70. https://doi.org/10.1207/s15430421tip4102_2