In recent years, the rapid development of digital technology has profoundly transformed the forms and methods of education. By breaking through the limitations of time and space, online education provides learners with more flexible and diverse learning options. For children aged 3 - 12, online education serves not only as a supplement to school education but also as an important extension of family education. Especially during special periods (e.g., public health events, severe weather, illness-related absences) and in special scenarios (e.g., remote areas, regions with weak educational resources), online education is a crucial means of safeguarding children’s right to education. However, online education for children aged 3 - 12 faces numerous practical difficulties. Unlike adult or adolescent learners, children in this age group are still developing in terms of attention span, self-control, abstract thinking, and other areas. Traditional “classroom migration” style online teaching often fails to achieve the desired results. Parents commonly report issues such as “children cannot sit still”, “easily distracted”, “insufficient interaction”, and “eye strain” ([14]). Some children even experience decreased learning motivation and poor learning outcomes. These problems are not inherent flaws of online education itself, but rather indicate that the design and implementation of current online education fail to fully respect children’s developmental patterns.

From a theoretical perspective, Piaget’s theory of cognitive development provides a classic framework for understanding children’s learning characteristics. Piaget divided children’s cognitive development into four stages: the sensorimotor stage (0 - 2 years), the preoperational stage (2 - 7 years), the concrete operational stage (7 - 11 years), and the formal operational stage (after 11 years) ([18]; [21]). Among these, the age range of 3 - 12 covers the preoperational and concrete operational stages, which are critical periods for children’s cognitive development. Children at these stages differ significantly from adults in terms of thinking characteristics, learning styles, and attentional features, and the design of online education must fully consider these differences.

This study aims to systematically analyze the practical challenges of online education for children aged 3 - 12, explore appropriate models for children’s online education based on cognitive development theory, and propose optimized pathways for curriculum design and teaching strategies. Using literature analysis and theoretical construction, the study seeks to provide a theoretical basis and practical guidelines for improving online education for children.

This study employed a theoretical synthesis approach rather than a systematic empirical review. To identify relevant literature, searches were conducted in the following databases: Web of Science, ERIC (Education Resources Information Center), PsycINFO, and the China National Knowledge Infrastructure (CNKI). The search terms used included combinations of the following keywords: “online education” OR “digital learning” OR “remote learning” AND “early childhood” OR “elementary school” OR “children aged 3 - 12” AND “cognitive development” OR “Piaget” OR “attention span” OR “multimedia learning”. The inclusion criteria were: peer-reviewed journal articles or books published in English or Chinese, studies focusing on children within the target age range (3 - 12 years), and works addressing at least one of the following domains: cognitive development, instructional design, or interaction patterns in digital environments. Priority was given to theoretical frameworks (Piaget, Vygotsky, Mayer) and empirical studies reporting age-specific attention data or interaction effects. The three theories were integrated by mapping Piaget’s stages to content difficulty, Vygotsky’s Zone of Proximal Development to scaffolding design, and Mayer’s principles to multimedia presentation ([19]; [16]). This three-dimensional framework (“Cognition-Society-Media”) was constructed deductively from the theoretical literature and then used to analyze practical challenges and derive design principles.

To ensure consistent application of the proposed framework, the following terms are defined at the outset:

Technology migration: The practice of directly transferring offline classroom models (e.g., 40-minute lectures, whole-class questioning) into online environments without adapting to the affordances or constraints of digital media or children’s developmental needs.

Development orientation: A design paradigm that takes children’s cognitive developmental laws (e.g., attention limits, concreteness of thinking, need for scaffolding) as the logical starting point for online education, with technology serving merely as a tool to achieve educational goals rather than as the goal itself.

Attention node: A pre-designed interruption or reset point inserted every 5 - 7 minutes within an online lesson (e.g., a question, a mini-game, or an action command such as “stand up”) intended to help children disengage from drifting thoughts and refocus on the learning task.

Para-social design: The incorporation of simulated social elements into online courses (e.g., virtual peer comments, animated characters that address the child by name, or pre-recorded teacher responses) that mimic real social interaction without requiring real-time peer presence. This term is adapted from para-social relationship theory to describe one-way, digitally mediated social cues.

Attentional problems are the primary challenge facing online education for children aged 3 - 12. Compared to offline classrooms, online learning environments contain more distractions: pop-up messages, household noise, other applications on electronic devices, etc. More importantly, children’s attention development has distinct age characteristics.

Research shows that the attention span of children aged 3 - 4 is approximately 5 - 10 minutes; 5 - 6 years, 10 - 15 minutes; 7 - 8 years, 15 - 20 minutes; 9 - 10 years, 20 - 25 minutes; and 11 - 12 years, 25 - 30 minutes ([15]). However, many current online courses still follow the 40 - 45 minutes class duration typical of offline settings, significantly exceeding children’s attention limits. This results in distracted attention during the latter half of the course and greatly reduced learning effectiveness.

Additionally, the reduced immediacy of teacher feedback in online teaching is another important reason for attention difficulties. In traditional classrooms, teachers can promptly bring distracted students back to attention through methods such as calling names, asking questions, or approaching them. In online environments, it is difficult for teachers to monitor each student’s attentional state, and once a child’s attention shifts, it often takes a long time to refocus.

Interaction is a crucial mechanism for children’s learning. Vygotsky’s sociocultural theory emphasizes that children’s learning occurs through interaction with others, and guidance from adults or more capable peers plays a key “scaffolding” role in children’s cognitive development ([19]; [22]).

However, current online education generally suffers from insufficient interactivity. A survey of online education platforms revealed that a considerable proportion of parents—for example, over 60% in one study—“insufficient interaction” a major drawback of online courses ([9]). It should be noted that this percentage is presented as an illustrative value to indicate the perceived prevalence of the issue, rather than as a precise estimate derived from large-scale statistical inference. This lack of interaction manifests in multiple dimensions: In terms of teacher-student interaction, it is difficult for teachers to ask questions frequently and provide immediate feedback in online classes, and students’ opportunities to ask questions and express themselves are also limited. In terms of peer interaction, common offline classroom interaction formats such as group discussions and cooperative learning are difficult to organize effectively in online environments. In terms of content interaction, static courseware and one-way video explanations lack interaction with learners, making it difficult to engage children’s sense of participation.

There are significant individual differences in the cognitive development of children aged 3 - 12. Among children of the same age and class, cognitive levels, learning styles, and interest preferences may vary considerably. In offline classrooms, experienced teachers can understand each student’s learning status through observation, questioning, and homework grading, enabling differentiated instruction. In online education, however, teachers face a screen behind which students are “indistinct”, making personalized guidance much more difficult.

Currently, mainstream online education models still rely primarily on “one-to-many” live or recorded sessions, where all students learn the same content, follow the same pace, and complete the same assignments. This “standardized” model overlooks individual differences, potentially leading to some children “not being challenged enough” while others “cannot keep up”.

Children aged 3 - 12 are in a stage of rapid physical development. Prolonged use of electronic screens may pose health risks such as vision decline, neck problems, and sedentary behavior. The “Technical Guidelines for Myopia Prevention and Control in Children and Adolescents” issued by the National Health Commission clearly states that children aged 3 - 6 should avoid using electronic products as much as possible, and primary and secondary school students should not use electronic products for more than 30 minutes per session ([14]).

However, the actual duration of online education often exceeds this recommendation. In addition to formal online courses, children may also use electronic devices after class to complete assignments or engage in recreational activities. Balancing learning effectiveness with screen time control is a practical issue that online education for children must address.

Compared to secondary school students, children aged 3 - 12, especially younger ones, require more adult support during learning. Online education shifts this support demand from school to the family. Parents need to help their children set up devices, log into platforms, maintain classroom discipline, assist with homework, and even accompany them throughout the course ([1]).

This “parental involvement” is a heavy burden for many families. Dual-income families may not have enough time to accompany their children throughout; in grandparent-headed households, grandparents may lack the skills to operate electronic devices; in multi-child families, parents find it difficult to simultaneously manage the online learning of several children. Some studies have pointed out that online education has, to some extent, exacerbated “parental anxiety” in education and has become a new source of burden on family education.

Piaget’s theory of cognitive development is an important theoretical tool for understanding children’s thinking characteristics. Piaget divided children’s cognitive development into four stages, with the age range of 3 - 12 primarily involving the preoperational stage and the concrete operational stage ([18]; [21]).

Preoperational stage (2 - 7 years): Children at this stage begin to use language and symbols for thinking, but their thinking remains markedly egocentric, one-dimensional, and irreversible. They have difficulty seeing things from others’ perspectives, tend to focus on a single dimension of an object, and cannot engage in logical reasoning. For example, in the classic “Three Mountains Task”, children in the preoperational stage cannot describe the view of the mountains from a doll’s perspective and can only describe their own viewpoint.

Concrete operational stage (7 - 11 years): Children at this stage begin to acquire logical thinking abilities, but logical operations still require the support of concrete objects. They can perform conservation reasoning (e.g., recognizing that the volume of a liquid remains the same despite changes in shape), classification, ordering, and other operations, but abstract reasoning abilities are not yet fully developed.

It is important to note that Piaget’s cognitive developmental stages are not discrete or sharply separated; rather, they overlap and transition gradually. Children aged 11 - 12 are in a transitional period between the concrete operational stage and the formal operational stage: some children begin to demonstrate preliminary abstract reasoning abilities, while others remain highly dependent on concrete support. Furthermore, the same child may exhibit different developmental levels across different knowledge domains (e.g., mathematics vs. language, science vs. art). Therefore, when applying cognitive stage theory to the design of online education for children aged 11 - 12, simplistic “labeling” should be avoided. Instead, stage characteristics should serve as a referential framework, with full attention given to individual differences and domain specificity.

Based on Piaget’s theoretical framework, the cognitive characteristics of children aged 3 - 12 and their implications for online education can be summarized as shown in Table 1:

3 - 4 years: Children in this stage are in the early preoperational period, with thinking highly dependent on concrete perceptions and marked egocentrism. They enjoy imitation and repetition, and have strong interest in bright colors and dynamic sounds. Online education should design very short sessions (within 10

Table 1. Cognitive characteristics of children aged 3 - 12 and implications for online education.

minutes), primarily using nursery rhymes, animations, and simple games, emphasizing repetition and imitation.

5 - 6 years: Language and symbolic abilities significantly increase. Children begin to understand simple storylines and can engage in role-playing. They enjoy pretend play and strongly identify with virtual characters. Online education can adopt forms such as picture-book stories and role-playing, with course duration within 15 minutes.

7 - 8 years: Entering the early concrete operational stage, children begin to have simple logical thinking abilities but still need concrete support. They enjoy hands-on manipulation, exploration, and discovery, and are interested in competitive activities. Online education should combine with hands-on activities (e.g., using blocks to learn math), keep course duration within 20 minutes, and introduce incentive mechanisms such as points and badges.

9 - 10 years: Further development of logical operations such as classification, ordering, and conservation. Children can understand more complex rules and are interested in cooperative activities. Online education can adopt forms such as project-based learning and group collaboration, with course duration within 25 minutes, encouraging peer interaction and presentation of outcomes.

11 - 12 years: In the later concrete operational stage, abstract thinking begins to emerge. Children can engage in preliminary hypothetical-deductive reasoning. They begin to pay attention to social issues, enjoy expressing their own opinions, and tend to question authority to some extent. Online education can introduce open-ended questions, debates, inquiry-based learning, and other forms, with course duration within 30 minutes, encouraging independent thinking and multi-perspective analysis.

Vygotsky’s sociocultural theory provides an important perspective for understanding the social nature of children’s learning. Vygotsky emphasized that children’s learning occurs in social interaction, and guidance from adults and peers plays a key role in children’s cognitive development ([19]; [20]).

Vygotsky’s concept of the “Zone of Proximal Development” (ZPD) points out that there are two levels of a child’s development: the “actual developmental level” (tasks the child can complete independently) and the “potential developmental level” (tasks the child can complete with help from others). The area between these two levels is the ZPD ([19]; [22]). Effective teaching should fall within the ZPD, neither below the child’s actual level (leading to boredom) nor exceeding the child’s potential level (leading to frustration).

The implication of this theory for online education is that online courses should not be one-way knowledge transmission; rather, they should be designed with ample interactive elements so that teachers and learning peers can provide “scaffolding” support for children. At the same time, course difficulty should dynamically adapt to the child’s ability level, providing appropriate challenge while remaining within the child’s reach.

A single theory cannot fully explain or guide the complex practice of online education. Based on Piaget’s cognitive development theory and Vygotsky’s sociocultural theory, this study further introduces Mayer’s cognitive theory of multimedia learning to construct a “Cognition-Society-Media” three-dimensional integrated framework suitable for online education for children aged 3 - 12 ([16]; [5]).

Mayer’s cognitive theory of multimedia learning posits that the human information processing system includes visual and auditory channels. Effective multimedia learning should avoid “channel overload” and promote meaningful learning through “relational processing”. Specifically, the theory proposes three core assumptions: the dual-channel assumption (visual and auditory information are processed separately), the limited capacity assumption (each channel can only process a small amount of information at a time), and the active processing assumption (learners construct knowledge by selecting, organizing, and integrating information) ([16]).

Integrating this theory with those of Piaget and Vygotsky yields the following design implications:

Cognitive dimension (Piaget): Course content and task difficulty should match the child’s cognitive developmental stage. For preoperational stage children, excessive use of abstract symbols should be avoided. For concrete operational stage children, logical tasks such as classification and ordering can be gradually introduced.

Social dimension (Vygotsky): Learning tasks should fall within the child’s ZPD and provide adequate “scaffolding” support, including teacher guidance, peer collaboration, and parental assistance.

Media dimension (Mayer): Multimedia presentation should follow the “redundancy principle” (avoid presenting text, animation, and narration simultaneously), the “coherence principle” (exclude extraneous information), and the “personalization principle” (use conversational rather than formal style) ([5]; [16]).

The proposed three-dimensional integrated framework transcends the explanatory limitations of any single theory and provides a more robust theoretical foundation for subsequent design principles and optimization strategies.

Children’s limited attention span is the primary fact that online education design must consider. The principle of short duration with high frequency requires breaking course content into multiple short units, each lasting within the child’s attention limit, and increasing learning frequency rather than extending single session duration, adopting a “small, frequent meals” learning rhythm.

Specifically, courses for ages 3 - 4 should be within 10 minutes; 5 - 6 years within 15 minutes; 7 - 8 years within 20 minutes; 9 - 10 years within 25 minutes; and 11 - 12 years within 30 minutes ([15]). In course design, a clear sense of “segmentation” should be established, with an “attention node” (e.g., questions, mini-games, action imitation) set every 5 - 7 minutes to help children refocus.

Play is the most natural way for children to learn. The gamification-driven principle requires embedding learning content into gamified tasks and scenarios, using game elements such as points, badges, leaderboards, and role-playing to stimulate children’s learning motivation ([13]; [24]).

Effective gamification design should satisfy three conditions: clear goals (children know what to do), immediate feedback (children know how they are doing), and appropriate challenge (difficulty matches ability). Research shows that gamified learning significantly improves children’s learning engagement and outcomes, especially for children aged 3 - 8 ([12]; [7]).

It should be noted that gamification does not mean “turning the course into a game”. The core of gamification is to stimulate intrinsic motivation, not to pile up external rewards. Over-reliance on material rewards or virtual points may backfire and diminish children’s interest in the learning content itself.

The thinking of children aged 3 - 12 is characterized by concreteness and imagery. Involvement of multiple sensory channels helps deepen understanding and memory. The multi-sensory participation principle requires that course design simultaneously engage children’s visual, auditory, kinesthetic, and other sensory channels, avoiding prolonged stimulation of a single channel ([4]).

For example, when learning numbers, children can look at number cards (visual), listen to number songs (auditory), trace numbers with their fingers (tactile/kinesthetic), and perform body movements related to numbers (e.g., clap five times). This multi-sensory integrated learning approach aligns with children’s cognitive characteristics and also helps reduce visual fatigue from prolonged screen viewing.

Online courses should encourage children to “get out of their seats” during learning—teachers can give action instructions such as “stand up”, “jump”, or “turn around” to get children moving. This “embedded movement” can alleviate health problems associated with prolonged sitting and also help maintain attention.

Learning is a social process, and peer interaction is important for children’s cognitive and emotional development. The social interaction embedding principle requires that online courses be designed with opportunities for teacher-student and peer interaction, avoiding a one-way “human-machine” model ([19]; [3]).

In terms of teacher-student interaction, teachers should frequently ask questions, invite students to share, and provide feedback on student work, creating a learning experience of “being seen”. In terms of peer interaction, activities such as group discussions, peer evaluation, and collaborative task completion can be set up to compensate for the lack of sociality in online learning.

For younger children, peer interaction may require assistance from adults (parents or teaching assistants); for older children, the grouping functions of online platforms can be used for small-scale interaction. Research has found that even very simple forms of interaction (such as showing each other’s work or taking turns speaking) can significantly enhance children’s sense of belonging and engagement.

Content modularization: Break course content into several independent yet interrelated modules, each focusing on one core knowledge point. Set clear transition signals (e.g., animated transitions, sound effects) between modules to help children develop a sense of “segmentation”.

Contextualized storytelling: Embed learning content in a coherent story line, allowing children to play roles in the story and advance the plot by completing tasks. Storytelling design can stimulate children’s imagination and immersion, reducing the sense of “task” in learning.

Difficulty gradation: Design learning tasks at different difficulty levels under the same topic, allowing children to choose an appropriate starting point and pace based on their level. Gradated design accommodates individual differences and provides children with a sense of challenge that is “within reach” ([6]).

Visualized evaluation: Children’s learning outcomes should be presented in intuitive, visual ways, such as a work display wall, progress curve chart, or skill tree unlocking. Visualized evaluation allows children to “see” their own growth, enhancing learning self-efficacy.

High-frequency interaction: Teachers should maintain a high frequency of questioning and feedback, avoiding lecture-style teaching. Research shows that initiating interaction (questions, polls, quizzes, etc.) every 3 - 5 minutes in online classes effectively maintains students’ attention ([3]).

Immediate feedback: Children require more immediate feedback than adults. Correct answers should be immediately reinforced with positive reinforcement (praise, points, etc.); incorrect answers should be met with gentle correction and hints, rather than a simple “wrong”. Immediate feedback helps children establish “action-outcome” connections and promotes learning adjustment ([8]).

Multimodal expression: Allow children to express learning outcomes in multiple ways, such as voice answers, drawing, photo uploads, video presentations, etc. Multimodal expression accommodates children with different expression preferences and provides teachers with richer information about student learning.

Emotional care: In online teaching, it is difficult for teachers to capture children’s emotional states through non-verbal signals. Therefore, teachers should consciously ask children “how are you feeling” and “are you happy today”, and intersperse relaxation, games, and other emotional regulation activities in the course.

Child-friendly interface: The platform interface should use child-friendly visual elements such as bright colors, rounded shapes, and cute icons. Buttons should be large enough for children to click easily, and operation processes should be simple, minimizing multi-level menus ([11]).

Simplified operation: Minimize children’s need for independent operation, using designs such as “one-click entry” and “auto-play”. For younger children, it is recommended that parents or teaching assistants assist with login, and children only need to participate in the learning session.

Eye protection mode: The platform should have a built-in eye protection mode (blue light filtering, automatic brightness adjustment) and set timed reminders. After a certain period of learning, an automatic “take a break” prompt should appear, guiding children to step away from the screen.

Offline availability: Considering that some families have limited internet access, the platform should support course download and offline learning. Offline functionality also facilitates learning in environments without internet access (e.g., travel, hospital rooms).

Parent guide: Provide parents with clear, easy-to-understand user guides, including device setup methods, course schedule explanations, and accompanying learning suggestions. Parent guides should avoid jargon and use illustrations supplemented by text ([1]; [10]).

Learning reports: Regularly push learning reports to parents, including learning duration, completion status, mastery level, interest preferences, etc. Learning reports help parents understand their child’s learning status and reduce “accompanying anxiety”.

Parent classes: Offer micro-courses for parents on topics such as child development knowledge, home learning environment setup, and parent-child communication skills. Parent classes are both a way to empower parents and a bridge to build home-school trust.

Flexible scheduling: Respect families’ different time schedules, provide recorded playback and flexible course time options. For children who cannot attend class on time for any reason, allow them to make up through playback and arrange targeted teacher guidance.

To better illustrate the necessity of optimization strategies, this study selects three common types of failed online courses for diagnostic analysis.

Classroom migration type: These courses directly record 40-minute offline classes as videos, with the teacher standing in front of a whiteboard lecturing, occasionally asking questions without actual waiting time. Typical performance: About 15 minutes into the course, students are visibly looking around, leaving their seats, playing with toys, etc. Diagnostic analysis: Violates the short duration with high frequency principle and the multi-sensory participation principle. Optimization direction: Break the course into 3 - 4 short units, each followed by an interactive game or action imitation session.

Over-animated type: These courses use a large number of dazzling animations, rapidly switching scenes, and exaggerated sound effects, attempting to attract children’s attention with high stimulation. Typical performance: Children are highly excited in the first 2 - 3 minutes, but attention drops sharply after 5 minutes, and they have vague memories of the course content itself. Diagnostic analysis: Violates Mayer’s multimedia learning “coherence principle” (too much extraneous information) and “limited capacity assumption” (visual channel overload) ([16]). Optimization direction: Reduce decorative animations, concentrate animation resources on presenting core concepts, and include a “pause and think” session after animations.

Isolated learning type: These courses are recorded, with no teacher-student or peer interaction at any point; children can only watch passively. Typical performance: After completing the course, children cannot recall the learning content, learning motivation continuously declines, and parents report that “children are unwilling to take it again”. Diagnostic analysis: Violates the social interaction embedding principle and overlooks the social learning mechanism emphasized by Vygotsky ([19]). Optimization direction: Add at least two forms of interaction—embedded questions (requiring children to answer orally or make gestures) and a work-sharing session (photo upload or voice presentation).

The summary of these three failed types provides a “pitfall avoidance guide” for course developers and makes the optimization strategies more targeted.

To move from theoretical principles to actionable evaluation, each of the four optimization dimensions (curriculum design, teaching strategies, technological adaptation, home-school collaboration) can be assessed using the following concrete metrics:

Curriculum design

Attention retention rate (proportion of children remaining on-task until the next attention node)Completion rate for modular unitsPre-to-post learning gain scores on modular assessments

Teaching strategies

Frequency of teacher-initiated interaction (questions, polls, action commands per 10 minutes)Student response rate (proportion of interaction prompts that receive a verbal, gestural, or typed response)Immediate feedback loop time (seconds between student answer and teacher/system feedback)

Technological adaptation

Average session duration before child-initiated exitNumber of operational errors per session (e.g., mis-clicks, navigation failures)Proportion of children completing a session without parental technical assistance

Home-school collaboration

Parent-reported accompanying time per session (minutes)Learning report viewing rate (proportion of parents who open the weekly report)Parent-reported confidence in supporting child’s online learning (pre-post Likert scale)

These indicators are proposed as a starting framework for design-based research. Future studies should establish reliability and validity for each metric across different age groups and subject domains.

Finland’s basic education is globally renowned for its high quality and equity, and in recent years, Finland has actively explored digital learning. Finland’s “phenomenon-based learning” emphasizes breaking down disciplinary boundaries and organizing interdisciplinary inquiry around real-world complex phenomena (e.g., “climate change”, “urban transportation”). In online education scenarios, Finland has developed the “Virtual Phenomenon Room” platform, allowing primary school students from different regions to observe experimental phenomena, collect data, and discuss solutions together through video conferencing. The key to the success of this model lies in its “slow-paced” design: each phenomenon unit lasts 4 - 6 weeks, with only 2 - 3 online synchronous sessions per week, leaving the remaining time for children to conduct offline observations and gather materials. This “online-offline deep integration” rhythm leverages the advantages of online connectivity while avoiding excessive screen time accumulation, demonstrating respect for children’s developmental patterns.

Singapore’s Ministry of Education launched the “Student Learning Space” (SLS), a national-level adaptive learning platform covering all levels from primary to high school. For the lower age group of 3 - 12, the SLS system specifically designed a “gamified adaptive engine” that dynamically adjusts the difficulty and presentation of subsequent content based on children’s real-time performance in subjects such as mathematics and language ([17]). For example, when the system detects that a child has made consecutive errors on a “fraction comparison” task, it automatically inserts a “visual fraction bar” manipulation tool and pushes relevant micro-lecture videos, rather than simply repeating similar questions. This “diagnosis-intervention-feedback” closed-loop mechanism essentially translates Vygotsky’s “Zone of Proximal Development” into algorithmic logic, providing personalized “digital scaffolding” for each child ([19]). Notably, while promoting adaptive learning, Singapore strictly limits online learning time for lower primary grades to no more than one hour per week and offers accompanying “digital literacy” parent workshops to ensure that technology application does not deviate from educational objectives.

Given its unique geographical environment, Japan has innovatively applied online education to children’s disaster prevention education. The “Bousai Mirai” project has developed a series of virtual reality (VR) experience courses, allowing children to simulate disaster scenarios such as earthquakes and fires in a safe virtual environment and learn emergency response skills. The educational design of this project fully considers Piaget’s stages of cognitive development ([18]): For preschool children in the preoperational stage (3 - 6 years), simple animated stories and role-playing are used, emphasizing action imitation of “follow what I do”. For primary school children in the concrete operational stage (7 - 12 years), situational decision-making tasks are introduced, such as “where should you hide during an earthquake?” and “how can you help an injured peer”, cultivating their preliminary problem-solving abilities. This case demonstrates that online education has unique value in achieving specific educational goals (such as safety education and health education), and its design should serve specific educational functions rather than pursuing technological showmanship.

The current challenges faced by online education for children aged 3 - 12 stem from an inertial thinking of “technology migration”—directly transferring offline classroom models online and replicating traditional teaching through technological means. This thinking overlooks the particularities of the online environment and, more importantly, the constraints that children’s developmental patterns impose on teaching formats.

This study proposes that effective online education for children requires a paradigm shift from “technology migration” to “development orientation”. Development-oriented online education takes the laws of children’s cognitive development as the logical starting point for design, with technology serving merely as a tool to achieve educational goals rather than being the goal itself. This means: course duration should follow attentional patterns rather than adult habits; interaction design should serve socio-developmental needs rather than merely “livening up the atmosphere”; technology adaptation should respect children’s physiological characteristics rather than pursuing feature accumulation.

This paradigm shift has implications for both educational technology research and practice. Researchers should pay more attention to the design and evaluation of “developmentally appropriate technology”, rather than merely pursuing technological “advancement”. Practitioners should return to the original educational purpose and adhere to the value orientation of “child-centeredness” amidst technological waves.

It must be emphasized that online education cannot and should not completely replace offline education. For children aged 3 - 12, offline education has irreplaceable advantages: face-to-face emotional communication, authentic peer interaction, rich sensory experiences, and timely physical feedback ([2]).

The relationship between online and offline education should be complementary, not substitutive. Online education has unique value in the following scenarios: emergency education during special periods (e.g., the pandemic); resource supplementation for remote areas; pathways for personalized learning; and educational support for special-needs children (e.g., children missing school due to illness).

Therefore, researchers and practitioners should rationally assess the functional boundaries of online education, neither blindly advocating nor completely rejecting it. Online education should be positioned as “an organic supplement to offline education” and “an integral part of the educational ecosystem”, rather than “a disruptor of offline education”.

This study primarily uses literature analysis and theoretical construction, lacking empirical data support. Future research can be conducted in the following directions:

First, conduct design-based research. Based on the design principles and optimization strategies proposed in this study, develop specific online courses and conduct small-scale experiments, collecting usage data and feedback from children, parents, and teachers to verify the effectiveness of the solutions and iterate improvements.

Second, compare the effects of different design elements. Use randomized controlled experimental designs to compare the effects of different course durations, interaction frequencies, degrees of gamification, and other factors on learning outcomes, providing empirical evidence for design decisions.

Third, track long-term effects. Conduct longitudinal follow-up studies to examine the long-term effects of online education on children’s cognitive development, learning habits, social skills, and other areas, assessing the sustainable value of online education.

Fourth, focus on special-needs children. Expand the research subjects from typically developing children to children with special needs (e.g., children with leukemia, ADHD, learning disabilities), exploring adaptive models of online education in special education contexts.

While affirming the value of online education, researchers should also maintain a prudent critical attitude. Based on existing literature, this study identifies the following four potential cognitive risks worthy of attention ([23]; [2]):

Risk of “fragmentation” of attentional habits: Although the high-frequency, short-duration format of online learning adapts to children’s attentional characteristics, it may also reinforce “short attention” patterns and weaken children’s ability to sustain focus on tasks requiring prolonged concentration (such as reading long texts or completing complex projects). This risk reminds us that online education should not be the only attentional training setting for children; offline deep learning remains a necessary supplement.

Risk of “substitution” of authentic social interaction: Although para-social design (e.g., virtual peer comments) is effective, over-reliance on it may reduce children’s opportunities for interaction with real peers, affecting the development of empathy and social cue recognition skills. The study suggests that para-social elements in online courses should not exceed 30% of total interaction time. It should be emphasized that this value is proposed as a suggestive threshold based on theoretical reasoning and design prudence, rather than as a precise cut-off derived from empirical evidence. The purpose is to caution designers against over-relying on simulated interactions at the expense of authentic social opportunities. The exact proportion should be adjusted dynamically according to age group and course objectives.

Risk of “atrophy” of embodied experience: Although the multi-sensory participation principle emphasizes action embedding, online learning is inherently disembodied. A child learning the concept of “apple” through a screen is qualitatively different from the embodied experience of actually touching, smelling, and tasting an apple. For children aged 3 - 6, online learning should not replace real sensory exploration activities.

Risk of “information cocoon” from algorithmic recommendation: Personalized recommendation algorithms may expose children to content that is highly matched with their interests and abilities for prolonged periods, reducing opportunities to encounter “unfamiliar” or “challenging” content, thereby limiting the breadth of cognitive development.

These risks do not constitute a rejection of online education but rather remind researchers and practitioners that the use of online education requires boundary awareness, balance awareness, and risk awareness.

Based on the full analysis, this study distills a ten-point practical quick checklist for frontline course developers:

Is an “attention node” (question, mini-game, action instruction) set every 5 - 7 minutes?Is storytelling, role-playing, or other gamification design used, rather than merely stacking points and badges?Are at least two sensory channels (visual, auditory, kinesthetic) engaged?Is teacher-student interaction (including embedded questions) initiated every 3 - 5 minutes?Are opportunities for peer interaction (group discussion, peer evaluation of work, collaborative tasks) designed?Does multimedia presentation avoid the redundant design of presenting “text + animation + narration” simultaneously?Are learning paths with different difficulty levels provided, allowing children to choose autonomously?Is an eye protection mode and timed rest reminder built in?Is a clear user guide and learning report provided for parents?Is course duration controlled within the recommended range for each age group?

This ten-point checklist translates the theoretical findings of this study into concrete, actionable, and checkable action items for course development.