Have a personal or library account? Click to login
A holistic evaluation framework for Chinese graded reading system Cover

A holistic evaluation framework for Chinese graded reading system

By: Rende Li and  Jian Zhang  
Open Access
|Oct 2025

Figures & Tables

Figure 1.

The graded reading evaluation criteria system. “Cost” and “Benefit” represent the direction of evaluation for each criterion. “Cost” refers to criteria where a lower value denotes easier reading difficulty. “Benefit” refers to criteria where a higher value denotes easier reading difficulty.
The graded reading evaluation criteria system. “Cost” and “Benefit” represent the direction of evaluation for each criterion. “Cost” refers to criteria where a lower value denotes easier reading difficulty. “Benefit” refers to criteria where a higher value denotes easier reading difficulty.

Figure 2.

The evaluation framework diagram. The process involves three steps from left to right: collect data, determine criterion weights, and calculate advantage degrees. Firstly, data is collected and aggregated using Probabilistic Fuzzy Linguistic Term (PFLT) and Probabilistic Linguistic Averaging (PLA). Qualitative data is assessed by experts using linguistic terms at five levels: Very Low (VL), Low (L), Medium (M), High (H), and Very High (VH), while quantitative data is directly measured values. Next, triangular and traditional entropy methods are used to calculate weights for the two criteria. Finally, the TODIM method calculates the advantage degree of each book based on the criterion and combines degree to rank the books.
The evaluation framework diagram. The process involves three steps from left to right: collect data, determine criterion weights, and calculate advantage degrees. Firstly, data is collected and aggregated using Probabilistic Fuzzy Linguistic Term (PFLT) and Probabilistic Linguistic Averaging (PLA). Qualitative data is assessed by experts using linguistic terms at five levels: Very Low (VL), Low (L), Medium (M), High (H), and Very High (VH), while quantitative data is directly measured values. Next, triangular and traditional entropy methods are used to calculate weights for the two criteria. Finally, the TODIM method calculates the advantage degree of each book based on the criterion and combines degree to rank the books.

Figure 3.

The comparison of book rankings across various evaluation methods. Different colored circles denote different books (B1, B2, …, B12). The vertical axis represents the ranking positions, while the horizontal axis represents the different evaluation methods. The lines of the same style track the changes in rankings for each book, highlighting the trends and variations in their ranking.
The comparison of book rankings across various evaluation methods. Different colored circles denote different books (B1, B2, …, B12). The vertical axis represents the ranking positions, while the horizontal axis represents the different evaluation methods. The lines of the same style track the changes in rankings for each book, highlighting the trends and variations in their ranking.

Figure 4.

The Kendall correlation between different ranking methods: WSTF’s Rank, Lexile’s Rank, Our Rank, Guide’s Rank, Exchange’s Rank, and Remove’s Rank. Significance levels are indicated as follows: *** p < 0.001 (two-tailed), ** p < 0.01 (two-tailed), and * p < 0.05 (two-tailed).
The Kendall correlation between different ranking methods: WSTF’s Rank, Lexile’s Rank, Our Rank, Guide’s Rank, Exchange’s Rank, and Remove’s Rank. Significance levels are indicated as follows: *** p < 0.001 (two-tailed), ** p < 0.01 (two-tailed), and * p < 0.05 (two-tailed).

Figure 5.

The comparison of ranking under different decision methods.
The comparison of ranking under different decision methods.

The comprehensive dominance, η(Bi) and ranking of all candidate books_

Benchmark booksComprehensive dominanceη(Bi)Ranking
B16.7141.0001th
B22.3250.9533th
B35.4010.9862th
B4-14.8150.7695th
B5-11.9590.7994th
B6-22.1230.6906th
B7-37.0060.5307th
B8-41.4610.4829th
B9-40.3410.4948th
B10-75.1070.12110th
B11-81.3140.05411th
B12-86.3590.00012th

Linguistic complexity and reading experience of the four Chinese classic novels_

Book titleLinguistic complexityReading experienceReasoning
Journey to the West (西游记)High difficultyLow difficultyLinguistic complexity: Rich in classical Chinese expressions, cultural references, and metaphors.
Reading experience: Linear and episodic storyline, vivid characters, and engaging plot make it easy to follow and enjoyable.
Romance of the Three Kingdoms (三国演义)High difficultyHigh difficultyLinguistic complexity: Complex sentence structures, historical terms, and strategic descriptions.
Reading experience: Dense historical and military content, with intricate relationships and strategies that require significant understanding.
Water Margin (水浒传)Medium difficultyMedium difficultyLinguistic complexity: More straightforward classical Chinese with less challenging syntax.
Reading experience: Many characters and subplots demand attention, but the heroic themes and action sequences are engaging.
Dream of the Red Chamber(红楼梦)High difficultyHigh difficultyLinguistic complexity: Elaborate classical Chinese with poetic and symbolic elements.
Reading experience: Complex emotional depth, subtle cultural references, and numerous characters and relationships make it demanding.

The entropy value and weight of each criterion_

CriteriaQuantitative criterion
C11C12C13C21C22C31C32C41
Entropy value0.5790.6020.6020.6840.7300.6930.6700.728
Initial weight0.1270.1330.1320.1500.1600.1520.1470.131
Normalized weights0.0760.0800.0790.0900.0960.0910.0880.072
Ranking8th6th7th3th1th2th4th10th

The existing graded reading systems_

Graded reading systemDescriptionLimitation
A-Z System (Hiebert & Tortorelli, 2022; McNamara et al., 2014)Categorizes books into 26 levels (A-Z), covering language difficulty and thematic content, with added factors like font and illustrations.Primarily designed for English; not easily adaptable to languages with different script systems.
Oxford Reading System (Gorard & See, 2016; Smith & Doe, 2018)Developed by Oxford University Press, uses vertical levels (based on age, cognitive and emotional development) and horizontal stages (e.g. phonics, comprehension).Structured for English-speaking readers; lacks accommodation for cultural and linguistic differences in other regions.
Developmental Reading Assessment (DRA) (Beaver & Carter, 2024; Johnson & Lee, 2017)A U.S. standard assessment evaluating reading comprehension, lexical knowledge, and reading strategies through progressive testing.Primarily assesses English skills; limited in flexibility for application to other linguistic contexts.
Lexile System (Hiebert, 2005; McNamara et al., 2014; Smith et al., 2016; Zeng & Fan, 2017)Uses semantic and grammatical complexity to determine reader levels and text difficulty, matching readers with suitable texts.Focuses on English text; lacks cultural adaptation for non-English readers.
Chinese Southern Graded Reading Center (Nur, 2019; Qiang et al., 2020)Based on the Lexile framework, divides grades 1-9 into four stages, considering text difficulty, narrative structure, and the integration of text and visuals.Primarily focuses on linguistic complexity; limited attention to reader interest and emotional engagement.
Shanghai Graded Reading Ability Standards (Holzknecht et al., 2022; Kidwai et al., 2016; Zhao, 2020)Adapts Lexile’s approach to measure reading attitudes, cognitive processes, and text difficulty in a Chinese context.Relies on linguistic complexity and overlooks personalized reading interests and emotional dimensions.
DOI: https://doi.org/10.2478/jdis-2025-0045 | Journal eISSN: 2543-683X | Journal ISSN: 2096-157X
Language: English
Page range: 197 - 218
Submitted on: Mar 24, 2025
|
Accepted on: Aug 10, 2025
|
Published on: Oct 16, 2025
In partnership with: Paradigm Publishing Services

© 2025 Rende Li, Jian Zhang, published by Chinese Academy of Sciences, National Science Library
This work is licensed under the Creative Commons Attribution 4.0 License.