Abstract

This study focuses on the digital transformation of primary school Chinese education in Macao, addressing issues such as superficial technology application and inadequate resource adaptability in traditional teaching. It proposes a dual-track model combining lightweight resource design with tiered dynamic assessment. Through tools like a modular digital resource library, real-time classroom feedback system, and electronic error portfolios, empirical research was conducted in eight primary schools in Macao. Results show that lightweight resources increased teaching efficiency by 37.2%, while the tiered assessment mechanism significantly improved students' text interpretation skills ( rate +28.5%) and writing transfer ability (+24.1%). This study provides actionable solutions for digitalizing Chinese education in Macao, supporting the development of an integrated "teaching-learning-assessment" ecosystem.

Keywords: teaching technology; digital resources; lightweight design; stratified assessment; Macao language education

1. Introduction: The Realistic Demands of Macao Language Digitization

As a cultural crossroads between China and the West, Macao's language education pursues dual objectives of humanistic cultivation and practical utility. Currently, digital teaching faces three major challenges:

Resource redundancy: High-definition videos and complex interactive resources cause slow loading, particularly in remote island school networks. In 45% of classrooms, resource retrieval delays exceed 3 seconds, severely disrupting teaching flow. For instance, a Taipa secondary school experienced 80% student feedback of screen lag during 3D classical poetry scene lessons due to bandwidth limitations, preventing real-time participation in interactive activities. Additionally, the existing resource library suffers from homogenization issues, with 62% of "Tang Poetry and Song Ci" themed resources being duplicated. Teachers spend an average of 12 hours weekly screening and adapting resources, resulting in low teaching efficiency.

Homogenized Assessment: Overreliance on standardized paper-and-pencil tests fails to capture higher-order literacy development (e.g., the artistic conception of "a waterfall plunging three thousand feet" in the poem "Viewing the Lushan Waterfall"). In traditional assessments, students 'understanding of "plunging" is limited to a 0.5-point multiple-choice question asking "What does this metaphor represent?" (A. water flow velocity B. steep mountain slope), without evaluating their ability to provide in-depth interpretations incorporating Li Bai's life and the grandeur of the Tang Dynasty. According to a 2023 survey by Macao's Education and Youth Affairs Bureau, 78% of Chinese language teachers believe current digital assessment tools cannot effectively measure students' critical thinking and cultural aesthetic abilities.

Feedback Lag: Test results are primarily used for grading and ranking, failing to provide timely instructional adjustments for over two weeks. For instance, after a third-grade "Expository Writing" unit test, teachers need three days to grade and five days to generate score analysis reports, by which time students have already moved on to the next unit, missing the optimal window for targeted tutoring. Data shows that classes using real-time feedback systems achieve a 40% improvement in students 'writing revision efficiency, compared to just 15% under traditional methods. Additionally, parental engagement with digital feedback remains low, with only 35% of families regularly reviewing their children's learning progress, which undermines the effectiveness of home-school collaboration.

Against this backdrop, there is an urgent need to establish a new paradigm for digital teaching that aligns with Macao's multilingual environment (Chinese, Portuguese, English) and prioritizes practical outcomes. This paradigm should integrate lightweight resource adaptation technologies (such as adaptive bitrate video), multidimensional literacy assessment models (combining AI sentiment analysis and text deep interpretation algorithms), and real-time feedback systems (supporting instant classroom diagnostics and personalized learning path recommendations). These innovations aim to overcome current bottlenecks in Macao's digital language education and achieve an organic integration of instrumental and humanistic elements.

2. Theoretical Basis: A Synergistic Framework of Lightweight and Hierarchical Evaluation

2.1 Principles of Lightweight Resource Design

Content refinement: Eliminate redundant animations while retaining core cognitive scaffolds. For instance, the "Borrowing Arrows with Straw Boats" courseware retains only the "Dynamic Route Diagram for Borrowing Arrows" and the "Character Relationship Mind Map" 1. Text resources employ segmented loading technology, with individual text resource packages ≤5MB3. According to the 2023 Education Informatization Development Report by the Ministry of Education, excessive animation in resources leads to a 37% distraction rate among students, whereas retaining core cognitive scaffolds improves knowledge retention by 42%. Technical compatibility

Compatible with legacy systems: PPT courseware automatically converts to H5 pages, ensuring compatibility with older tablet models. Offline resource packages: Pre-installed classical poetry recitation audio files on classroom terminals mitigate network instability. A survey at a Macau secondary school revealed that implementing these offline resources reduced teaching disruptions caused by network issues by 89%, significantly enhancing educators' instructional efficiency.

2.2 Hierarchical Dynamic Evaluation Model

Based on the fourth-generation evaluation theory, the "three-dimensional evaluation system" is constructed:

A[Process Diagnosis] --> A1 (In-Class Immediate Feedback)

A--> A2 (Electronic Error Record File)

B[Contextualized Task] --> B1 (Practical Assessment)

B → B2 (cross-media expression)

C[Ability Advancement] --> C1 (Hierarchical Achievement Scale)

C → C2 (Learning Path Map)

3. Practice Case: Application of Localized Classroom

3.1 Application of Lightweight Resources in Teaching of Ancient Poetry and Prose

3.1.1. Classical Chinese Decoding Toolkit

The project designed a "Comparative Matrix of Ancient and Modern Meanings" for twelve passages from The Analects, requiring annotation of the phonetic loan character "shuo" in "Bu yi shuo hu" (Isn't it delightful?), along with its lexical application in "Yu lü di xi" (crouching and carrying) from The Record of the Drunken Old Man's Pavilion (4). The lightweight implementation replaced semantic animations with static comparison tables, reducing resource size by 78% (3). Tests demonstrated that when paired with teacher explanations, this static table increased students' accuracy in mastering classical Chinese function words from 65% to 89%, while tripling the resource loading speed.

3.1.2. Micro-resources for Reading Practice

Develop the "Three-Step Reading Module":

Base layer: Mark the pause and pause symbols of "Thoughts on a Tranquil Night"

To the Class: Contrast of Emotional Differences in Cantonese/Mandarin Reading

Create Layer: Create a soundtrack for "Moonlight" and explain the reason

The offline audio package covered 10 primary schools in Macao, achieving a utilization rate of 92%. Among the creative layer tasks, students submitted 127 music composition proposals, encompassing various genres such as classical instruments and modern pop music, demonstrating their rich cultural understanding and innovative expression capabilities.

3.2 Classroom Implementation of the Hierarchical Assessment System

3.2.1. Design of Layered Tasks (Take "Bird Paradise" as an Example)

3.2.2. Electronic Error Portfolio Function

Intelligent Attribution Analysis: The system employs natural language processing technology to conduct in-depth analysis of students 'incorrect answers, automatically categorizing error types and generating detailed error classification statistics. For instance, "Rhetorical Confusion" errors account for 35.7%, primarily manifesting as students' difficulties in identifying and applying rhetorical devices such as metaphors, personification, and parallelism. Examples include misidentifying "metaphor" as "personification" or failing to accurately grasp the core characteristics of rhetorical devices when imitating sentences. "Contextual Misreading" errors make up 28.1%, predominantly occurring in reading comprehension or cloze tests. These errors arise when students fail to accurately interpret logical relationships, emotional tones, or implicit information in textual contexts, leading to distorted understanding of word meanings and sentence themes. A common example is interpreting irony as literal meaning. Additionally, the system identifies other prevalent error types, including "Grammatical Structure Errors" (18.3%), "Inappropriate Word Pairing" (9.2%), and "Logical Reasoning Errors" (8.7%). Through visual charts, the system visually displays the distribution of various errors, helping students and teachers quickly pinpoint weak learning areas.

Error Correction Path: The system provides personalized error correction resources and learning paths tailored to different types of errors. Taking the typical grammatical error of "de, di, de" mix-up as an example, when students make such mistakes during practice, the system immediately triggers the correction mechanism and pushes targeted learning resources. Specifically, the system delivers 1-2 minute animated micro-lessons titled "Part-of-Speech Identification Mnemonics" (file size ≤2MB, ensuring fast loading on all devices). The animation vividly explains how "de" modifies nouns as attributives, "di" modifies verbs as adverbs, and "de" supplements actions with results or states before complements. It also demonstrates comparisons through numerous example sentences like "run quickly,"  "beautiful scenery," and "run fast." Additionally, the system provides 5 instant practice questions for students to apply their knowledge immediately after watching the animation, reinforcing learning outcomes. For students with higher error rates, the system further offers advanced exercises, such as analyzing complex sentence structures involving "de, di, de" mix-ups and specialized training for easily confused words. This creates a complete error correction loop of "error identification-resource delivery-instant practice-feedback," effectively improving students' grammatical accuracy.

4. Empirical Effects and Teaching Reflection

4.1 Quantitative Data Validation

Conduct comparative experiments at eight schools including Haojiang Primary School and Peizheng School (September 2025 to January 2026):

Table 1 Effectiveness of Lightweight Resource Application (N=426)

Table 2 Stratified Assessment Outcomes (Writing Specialization)

(2) Qualitative feedback

Teacher: 'The digital error portfolio precisely identifies the "ge-dé" (geographical and linguistic) confusion among students in remote islands. After implementing Cantonese dialect analysis nursery rhymes, the error rate dropped by 60%' (Ms.Huang, St. Joseph Parish Primary School).

Student: 'When composing music for "Drinking on the Lake: First Clear, Then Rain," I chose a Portuguese guitar piece. The teacher said this is called cultural fusion' (Gianosan Sacred Heart English Department P5B).

(3) Existing Challenges and Countermeasures

The gap in teachers' technical literacy

Problem: Only 31% of teachers at remote island schools are proficient in resource compression techniques, resulting in excessive file sizes for lesson resource packages (averaging over 20MB). This causes slow downloads in the limited bandwidth of remote island networks, severely impacting teaching efficiency. Additionally, about 68% of teachers report difficulties using lightweight tools during resource processing, further exacerbating technical application barriers.

Solution: The Education Youth Bureau has developed the "Lightweight Toolkit 1.0", featuring a step-by-step tutorial on format factory, a 7-Zip compression software guide, and a repository of resource optimization cases. Tailored to the technical proficiency of teachers in remote islands, the toolkit employs a dual-mode approach combining visual aids and short videos. Three specialized training sessions have been conducted, reaching 120 teachers. Initial feedback indicates a 40% improvement in resource compression efficiency, with average package sizes now under 8MB.

Barriers to cross-platform compatibility

Problem: The incompatibility between the e-portfolios system and the school-based academic management system's data formats prevents direct import of students' learning process data (e.g., homework submission records, classroom performance, etc.). Teachers must manually enter this data, which takes approximately 2 hours per class and is prone to errors. According to a survey, 85% of teachers believe this issue affects the timeliness and accuracy of teaching evaluations.

Solution: Establish a CSV universal data interface to standardize the data field specifications for electronic portfolios (including 12 core fields such as student ID, academic number, date, grades, and comments), supporting batch export and import across systems. System integration testing with three off-island schools has been completed, achieving a data import accuracy rate of 98.5%, while reducing teacher data processing time to 15 minutes per class.

5. Conclusion: Establishing a Macao-specific digitalization pathway

Resource Optimization Strategy: Adopting "content refinement" over "technical redundancy", we develop lesson resource packages under 5MB by compressing non-essential audiovisual elements and optimizing image resolution to ensure lightweight resources without compromising teaching effectiveness. For example, a 45-minute Chinese language lesson package was compressed from 25MB to 4.8MB while retaining key knowledge points, blackboard screenshots, and interactive design. A "Island-Peninsula" resource synchronization mechanism was established, pre-synchronizing Peninsula's premium resources to local servers during off-peak nighttime hours via dedicated lines to overcome bandwidth limitations, reducing average resource access response time from 30 seconds to under 2 seconds.

Reassessing Teaching Approaches: Implementing "Tiered Task Cards" designs differentiated learning tasks based on students' Chinese proficiency levels (Basic, Advanced, and Extended tiers). The Basic tier focuses on vocabulary building and sentence pattern imitation, the Advanced tier emphasizes reading comprehension and writing expression, while the Extended tier prioritizes literary appreciation and critical thinking, achieving "one-person-one-strategy" precision teaching. Enhancing the diagnostic function of digital portfolios, the system automatically identifies learning weaknesses through data analysis (e.g., high error rates in ancient poetry recitation, illogical responses in modern text reading) and generates personalized diagnostic reports. Targeted intervention resources (e.g., specialized exercises, micro-lecture explanations) are then recommended, forming a closed-loop management cycle of "diagnosis-intervention-reassessment." Pilot data shows that classes adopting this model achieved a 12% improvement in average Chinese scores and significantly enhanced learning initiative.

The vision of ecological synergy: Lightweight resources streamline teaching by reducing burdens and boosting efficiency, freeing teachers from tedious technical tasks to focus on instructional design and student mentoring. Tiered assessments provide precise learning navigation, helping students clarify goals and pathways to enhance learning outcomes. Together, these initiatives propel Macao's language education from "technology application" to "ecological reconstruction," establishing a student-centered, data-driven, and lightweight resource-based digital teaching ecosystem that fuels high-quality educational development in Macao.

References

[1]. Macao Education and Youth Development Bureau. Macao Non-Higher Education Digitalization White Paper [R]. 2025:23-27.

[2]. Chen Xiaohui. Research on the Construction of a Digital Education Evaluation System [J]. China Electro-Education, 2025(4):67-73.

[3]. Lin WM. Design of Lightweight Teaching Resources in the Context of Mobile Learning [J]. Journal of Distance Education, 2025(1):55-62.

[4]. Huang Wanting. Exploring Digital Teaching Strategies for Primary School Chinese [J]. Educational Research and Experiment, 2024(6):88-92.

[5]. Zhu Jinxin. Application pathways of process evaluation in Chinese language teaching[J]. Curriculum, Textbook and Teaching Method, 2025(3):112-118.

[6]. Macao Diocesan School Federation. Research Report on Information Technology Application in Island Schools [Z]. 2025.