Sustainable rural development: differentiated paths to achieve rural revitalization with case of Western China

Sustainable rural development is vital for society by providing essential raw materials and supporting stable growth in various national industries^1,2. Furthermore, rural areas facilitate urban development by offering affordable labor, a robust consumer market, and a conducive atmosphere for urbanization^3,4,5. Additionally, rural regions enhance national security through their independent economic systems and rich natural resources; they serve as strategic buffer zones during crises, safeguarding the country’s food supply and land⁶.

However, sustainable rural development currently faces several challenges^7,8. First, migration to urban areas has resulted in agricultural labor shortages⁹; second, the low-added value of agricultural products fails to boost villagers’ income and sustainable rural economies¹⁰; third, increased demand for non-agricultural buildings occupy much arable land, hindering agricultural development¹¹; fourth, rural areas lack adequate infrastructure and services like roads, water supply, power supply, and healthcare compared to cities, limiting villagers’ quality of life¹²; fifth, excessive pesticide and fertilizer use degrades soil quality and harms the rural environment¹³; sixth, aging populations pose governance challenges in rural societies¹⁴; finally, preserving rural culture is difficult due to brain drain affecting practices and intangible heritage¹⁵. These issues significantly impede sustainable rural development and highlight the urgent need for solutions.

The Sustainable Development Goal (SDG) effectively addresses the above rural challenges by promoting inclusiveness, safety, disaster resilience, and sustainability^8,16. Specifically, it ensures agricultural production, essential services, safe community management, cultural heritage preservation, environmental protection, disaster loss reduction, and improved quality of life for rural residents¹⁷. Many regions have made progress in this topic under the guidance of SDG. For instance, Europe’s Common Agricultural Policy protects villagers’ rights while enhancing food security and productivity to boost the rural economy¹⁸. France focuses on improving life quality and countryside attractiveness; Germany emphasizes ecological agriculture and digital initiatives for village modernization^19,20,21. Developing countries have also made progress; for instance, China’s Rural Revitalization (RR) strategy aims at sustainable development through five key goals in villages: industrial prosperity, ecological livability, civilized customs, effective governance, and enriched lives²². It seeks coordinated and environmentally friendly growth in rural areas, making it a valuable subject for further research.

Scholars have theoretically explored RR, the practice of sustainable rural development in China. Some studies discuss its prerequisites, emphasizing external support through urban–rural integration and policy assistance, rural self-reliance, and integration of internal motivation and external support^23,24. Others highlight RR’s importance: it optimizes rural economic and social structures, fosters fairness and harmony, improves community well-being, preserves natural resources, and benefits local populations²⁵. Some research also examines measurement frameworks like “5-E,” “social-ecological,” and “element-structure–function,” along with indicators such as “electricity consumption per capita” and “urban–rural income ratio” to evaluate RR performance^8,26,27. Additionally, RR’s challenges include a lack of distinct rural characteristics, insufficient self-motivation for development, weak governance, and resource loss^28,29. Corresponding solutions proposed include enhancing urban–rural interactions, providing policy support, encouraging regional collaborations, promoting industrial innovations, improving public services, conducting human resource training, and fostering social participation^30,31,32. These measures aim to assist rural areas in achieving RR.

However, current RR research raises the following questions, which are also the objectives of this study:

1. (1)

   Is there a clear and explicit RR framework that simplifies the global theoretical understanding of RR? (This will be analyzed in section “3.1. Framework Establishment.”).

2. (2)

   Is there an evaluation system with easily understandable indicators for widespread application? (This will be analyzed in section “3.2. Indicator System.”).

3. (3)

   What is the current status and differences of RR based on the framework (research question 3–1; to be analyzed in “4.2. Analysis of Overall Development Status”), and how can various villages enhance aspects to achieve RR (research question 3–2; to be analyzed in “4.3. Differentiated Paths of RR”)?

4. (4)

   Are there commonalities in the various paths to implementing RR (research question 4.1; to be analyzed in “4.4. Categories of Differentiated Paths”), and what specific actions can villages in different contexts take to achieve RR (research question 4–2; to be analyzed in “4.5. Countermeasures”)?

This study conducts the following details to address the above corresponding questions:

1. (1)

   To achieve the first objective, a new “5I Framework” is developed, consisting of five dimensions starting with "I," making it easier to understand and disseminate globally³³.

2. (2)

   Based on this framework, a new indicator system with globally representative indicators for broader usability will be created to achieve the second objective.

3. (3)

   To achieve the third objective, the overall development status and the differentiated paths for improving it in achieving RR are analyzed, identifying specific aspects that various villages can enhance.

4. (4)

   To achieve the fourth objective, the paths are categorized into “4L Category,” named for their initial letters to facilitate understanding. Similar villages can adapt implementable strategies from each category to effectively achieve RR.

The section divisions are as follows. Section 2 reviews relevant theories and establishes the theoretical foundations for classifying RR dimensions. Section 3 outlines the framework, methods, and data; it introduces the “5I Framework” with indicators to evaluate RR, selects appropriate methods and data, and details the calculation process. Section 4 discusses results and offers suggestions; using the 5I Framework, this study analyzes RR’s overall development status (ODS) while comparing temporal and spatial differences. Section 5 highlights core results, novelties, limitations, and future research directions.

The study has several innovations. Firstly, it presents a new 5I Framework that encompasses various dimensions influencing RR and establishes a comprehensive evaluation system tailored to the framework, enhancing theoretical research on RR and contributing new insights into RR to the existing literature. Secondly, integrating the entropy weight method, the technique for order preference by similarity to an ideal solution, and fsQCA provide objective, replicable, and valid methodological references for other studies in other contexts. Thirdly, the study examines regional RR’s temporal and spatial dynamics, providing readers with a dynamic overview of current trends. Fourthly, it identifies differentiated RR paths across regions, categorizing them into “4L Category” and suggesting targeted countermeasures to support each category’s RR. Notably, the “5I” Framework consists of five dimensions; through the fuzzy set qualitative comparative analysis (fsQCA) calculations, distinct paths to achieve RR are identified and clustered into four categories based on the paths’ similarities.

Literature review

Theoretical basis

Rural revitalization (RR) addresses three key rural issues: agriculture, villagers, and the countryside³⁴.Firstly, agriculture is the backbone of rural areas and essential for RR; resolving agricultural challenges is crucial for sustainable rural development³⁵. Secondly, villagers are vital to agricultural production and beneficiaries of RR; improving their quality of life and happiness is a primary goal³⁶. Thirdly, the countryside supports both agriculture and villager growth while preserving rich cultural heritage and natural resources that facilitate RR progress³⁷. Therefore, exploring differentiated paths in RR requires considering these aspects (agriculture, villagers, and the countryside) for comprehensive coordination.

RR requires the involvement of various stakeholders. It goes beyond simply applying urban development models to rural areas; instead, it represents a shift towards sustainable rural development by utilizing inherent resources and external support from different participants³⁸. Firstly, key components of RR include government support, which offers policy and financial assistance while encouraging enterprise involvement³⁸. Secondly, active participation from enterprises and society is vital; businesses can integrate local resources and upgrade traditional agriculture via new technologies, while social organizations raise public awareness about RR^14,39. Thirdly, villages themselves are crucial; they must recognize their unique advantages to achieve differentiated positioning and pursue distinct development paths like agritourism or elderly healthcare services^40,41,42. Fourthly, villagers are important; their active engagement is essential for shifting from reliance on social welfare to becoming contributors to RR^43,44. Fifthly, professional agricultural producers are needed; robust agricultural development is crucial for the overall functioning of society⁴⁵.

Correlations between RR and Its goals

Studies indicate that RR involves factors like agricultural production, ecological environment, culture and education, infrastructure development, and villagers’ economic income^46,47,48. China’s RR aims for five goals (the Chinese government calls the overall requirements): industrial prosperity, ecological livability, civilized rural customs, effective governance, and improved life quality²². These goals reflect diverse aspects of sustainable rural development; analyzing their correlation with RR helps establish a comprehensive theoretical framework.

Basis of RR: industrial prosperity

The prosperous agricultural industry is the foundation of RR. Firstly, agriculture is the core industry in rural areas, satisfying villagers’ essential material needs⁴⁹. Secondly, increased agricultural productivity and efficiency from mechanization boost villagers’ incomes and facilitate the transfer of surplus labor^50,51. Thirdly, enhancing agriculture’s resilience to disasters is crucial for high-quality RR because improved recovery capabilities after catastrophes promote more agricultural outcomes³⁶. Overall, a thriving agricultural sector is essential for sustainable rural development.

Support of RR: ecological livability

Ecological livability is a key pillar of RR. Firstly, a healthy ecological environment fosters an environmentally friendly and hygienic living space, attracting residents and promoting rural development, accelerating RR⁵². Secondly, good rural ecological conditions support agricultural growth by effectively monitoring and controlling pollution from farming practices, leading to sustainable rural production⁵³. Thirdly, essential infrastructure in rural areas enhances living and working conditions; improved water and sewage facilities boost residents’ well-being and satisfaction, encouraging active participation in RR⁵⁴.

Safeguard of RR: civilized rural customs

Civilized rural customs are crucial in developing rural culture and education, serving as a spiritual foundation for RR. Firstly, improved rural education boosts agricultural productivity; educated villagers can quickly identify production issues and adopt new techniques⁵⁵. Secondly, better education empowers villagers to engage effectively in grassroots management, promoting sustainable countryside development¹⁴. Thirdly, civilized customs and culture ensure that RR remains up-to-date; modern technologies like the internet facilitate the evolution of these customs, helping villagers discard outdated values and embrace new ones that align with contemporary society⁵⁶.

Accelerator of RR: effective governance

Effective governance enhances the efficient use of public resources and policies for RR. Firstly, the government can mobilize substantial social resources to invest in rural infrastructure, including health services, roads, and essential utilities, supporting RR^57,58. Secondly, effective governance fosters social security, stability, and equity in rural areas. Local governments allocate funds to assist low-income villagers, ensuring that RR benefits are shared among all community members and promoting harmony within villages^59,60. Thirdly, effective governance encourages democratic participation in rural areas and adherence to laws and local regulations. That leads to the sustainable and harmonious development of RR⁶¹.

Core of RR: improved life quality

The improved quality of life for villagers is a crucial aspect of RR. Firstly, wealthier villagers boost rural consumption and economic growth; those with stable incomes are more likely to spend on entertainment, stimulating the development of other service industries⁶². Secondly, improving low-income villagers’ earnings through coupons and targeted poverty alleviation reduces income inequality, fostering a sense of dignity and motivating them toward sustainable rural development⁶³. Thirdly, enhancing villagers’ consumption status enables upgrades in agricultural production and community management facilities, benefiting RR in agriculture and governance⁵⁹.

These five goals are interconnected and focus on agriculture, countryside, and villagers. Each goal plays a unique role in RR. However, villages differ in characteristics; thus, these goals have varying impacts on RR outcomes for different villages. It is essential to explore how they collectively influence the path of RR under specific scenarios—specifically, how different villages achieve RR from these five perspectives. This exploration helps scholars understand diverse RR paths and identify paradigms for sustainable rural development.

The above analysis revealed complex correlations among industrial prosperity, ecological livability, civilized rural customs, effective governance, improved life quality, and RR. Therefore, a framework for evaluating RR can be established based on these goals.

Framework establishment

To facilitate differentiated RR path analysis, it is essential to construct a clear and concise framework for simplifying complex multi-factor issues. Previous studies have proposed various frameworks for evaluating rural revitalization or development. This study conducts a bibliometric analysis to explore RR frameworks with attributes. This study collected English articles published between January 1st and December 1st, 2024, containing the terms {“rural revitalization” or “rural develop*”} in their titles. Articles indexed in the Web of Science Core Collections, including the Social Sciences Citation Index and Science Citation Index Expanded, were specifically selected due to their peer-reviewed status and high recognition. Two members independently searched and screened out papers; a third member resolved any disagreements. Initially, we identified 138 papers; after screening out irrelevant ones (e.g., those not mentioning RR frameworks or retractions), we finalized 42 articles. Detailed bibliometric results are presented in Table 1. Notably, five dimensions—industry, ecology, customs, governance, and life quality—are commonly emphasized. However, although these frameworks convey the essence of RR, they can be hard to remember.

This study presents an innovative RR theoretical evaluation framework that is easy to understand and remember, answering research question 1: “Is there a clear and explicit RR framework that simplifies the global theoretical understanding of RR?” and aiding subsequent analysis. The framework includes five dimensions representing the goals of RR: “Industry (INDUS),” “Inhabited environment (INHAB),” “Indoctrination (INDOC),” “Infrastructure (INFRA),” and “Income (INCOM).” Collectively, these form the “5I Framework” based on their initial letters, as illustrated in Fig. 1. The 5I Framework simplifies understanding complex RR goals with its precise dimensions; thus, the 5I Framework represents a significant improvement in remembering and sharing.

5I Framework.

Full size image

We want to highlight that “dimensions” in the framework and “goals” in RR are different while correlated. They are different because the goals in RR are proposed by the Chinese government (also called overall requirements), while the names of dimensions in the framework are proposed in this study. Meanwhile, they are correlated because the goals in RR correspond to the dimensions in the framework; in other words, the dimensions are derived from the goals. The goals of “industrial prosperity,” “ecological livability,” “civilized rural customs,” “effective governance,” and “improved life quality” correspond to the dimensions of INDUS, INHAB, INDOC, INFRA, and INCOM, respectively.

INDUS dimension

INDUS highlights the vital role of “agricultural industry growth” in RR and has the following connotations. Firstly, continuous agricultural growth ensures stable incomes for villagers and supports other rural sectors; a thriving agriculture sector is essential for RR⁷². Secondly, agricultural mechanization significantly boosts productivity and output value, enhancing RR performance⁵⁰. Lastly, protective measures for the agricultural industry are crucial as they mitigate natural disasters, ensuring stable rural development and RR⁸.

INHAB dimension

INHAB highlights the essential “living conditions of villagers” in the RR process with the following connotations. Firstly, adequate water and sanitation facilities are crucial for maintaining these conditions⁴. Secondly, increased green coverage is important for improving the living environment and promoting RR⁷³. Thirdly, agricultural practices can lead to air and water pollution (e.g., pesticide use); hence, a better living environment indicates that measures have been taken to mitigate these effects, ensuring villagers can live and work in an eco-friendly countryside⁷⁴.

INDOC dimension

INDOC reflects the “educational status of the rural population”, including educational resource allocation and information exchange. Firstly, rural educational resources, such as community cultural centers, are vital for rural development; investing in these resources improves RR^55,75. Secondly, rural teachers play a crucial role; exceptional educators cultivate capable villagers and enhance the local cultural atmosphere⁷⁶. Thirdly, modern technology like the internet is essential for informing villagers about national strategies, thereby boosting RR effectiveness⁷⁷. An improved educational and cultural status enhances villagers’ understanding of RR, fostering a more coordinated approach to developing RR⁴⁶.

INFRA Dimension

INFRA involves managing “public resources and infrastructure in rural areas”, such as healthcare, roads, electricity, and poverty support. Firstly, building rural infrastructure is essential for RR; it promotes harmony across various RR aspects and secures residents’ support for government initiatives^43,78. Secondly, public services like protecting low-income individuals enhance stability and social security in the countryside⁷⁹. Thirdly, investing in rural infrastructure boosts the local economy, fosters orderly industrial development, and ensures the systematic progress of RR⁸⁰.

INCOM dimension

INCOM reflects “villagers’ income and expenditure conditions”, including income equality, quality of life, and consumption status. Firstly, relative income equality ensures villagers’ satisfaction and well-being, motivating them to pursue better RR⁸¹. Secondly, increased income and improved quality of life are crucial; they indicate RR outcomes, heighten demand for a better life, and facilitate further RR efforts⁸². Thirdly, stable living expenditures are vital; they stimulate rural economic growth, maintain social stability in rural areas, and enhance acceptance of RR⁸³.

In summary, the 5I Framework depicts essential components of RR, working together in specific combinations for differentiated RR growth. It offers several advantages: (1) Global applicability: The dimensions are universally representative and serve as a paradigm for sustainable rural development research. (2) Intuitive understanding: Named with initial letters, the framework clearly emphasizes RR’s essence—addressing three rural issues: agriculture, villagers, and countryside—facilitating differentiated path analysis. (3) Comprehensive perspective: It captures the coordinated interactions of RR dimensions, providing a dynamic view of complex elements without oversimplifying the process.

The “5I” is a theoretical evaluation framework that depicts every dimension’s connotation and assists villages in identifying their unique development paths. It balances simplicity and comprehensiveness to effectively address complex issues, affirming research question 1: “Is there a clear and explicit RR framework that simplifies the global theoretical understanding of RR?”.

Indicator system

Indicator selection

After establishing the 5I Framework, it is essential to create a scientific indicator evaluation system based on its connotations to analyze the differentiated path of RR⁸⁴. Drawing from previous research^85,86, this paper develops a comprehensive and representative indicator evaluation system guided by the following principles, addressing question 2: “Is there an evaluation system with easily understandable indicators for widespread application?”.

• (1) Indicators reflect core ideas and connotations related to RR dimensions.

• (2) Indicator data are easily accessible and officially referenced.

• (3) Indicators must be objective and unbiased, free from human intervention.

• (4) Indicators should have universality for global applicability in sustainable rural development.

• (5) Indicators aim to reflect “per capita” metrics for better comparability and reduced errors due to regional differences.

• (6) Indicators must pass covariance tests.

• (7) Indicators should be easily understandable.

• (8) Indicators have been widely used in prior studies and recognized as valid.

The indicator selection process is as follows, illustrating why specific indicators are preferred over others.

• (1) Two team members independently review the literature to identify relevant indicators and their variants.

• (2) They screen the indicators based on established criteria, favoring “per capita” metrics when similar options exist and excluding those lacking cross-regional longitudinal data.

• (3) The members cross-check the selected indicators; common choices are made, and a third member intervenes in case of disagreements for an independent decision.

• (4) The shortlisted indicators are sent to two experts in rural revitalization for feedback, leading to the final indicator system after expert review.

These procedures ensure relative objectivity in selecting indicators.

In the end, 26 indicators were chosen (Tables 2–6). In this study, all indicators are classified as positive or negative according to their impacts on RR and previous studies. Positive indicators are indicated with “ + ,” signifying a favorable effect on RR, and negative indicators with “-,” indicating negative effects on RR.

Indicators of INDUS

The indicators are in Table 2. Agriculture is the primary industry and income source for most rural residents, directly impacting the development of other sectors and promoting RR. This study uses six indicators to assess agricultural development: proportion of agriculture-added value in gross regional product, agricultural labor productivity rate, grain output per capita, total power of agricultural machinery per capita, area of cultivated land plowed by machinery per capita, and proportion of natural-disaster-affected crop area to total sown area. These indicators provide a comprehensive evaluation of the agricultural sector’s efficiency, productivity, and mechanization.

Formula: INDUS 1 = agriculture-added value / gross regional product of the year; INDUS 2 = total power of agricultural machinery / rural population; INDUS 3 = grain output / rural population; INDUS 5 = area of cultivated land plowed by machinery / rural population; INDUS 6 = natural-disaster-affected crop area / total cultivated land area. Data for these indicators can be directly obtained from relevant yearbooks. Indicators without a formula indicate that data is also available in the corresponding statistical yearbook, and this applies to all similar cases.

Indicators of INHAB

The indicators are in Table 3. The rural living environment, a key goal of RR, must address ecological and living conditions. This study uses five indicators: water supply penetration rate in houses, per capita drainage pipe length, toilet coverage ratio in houses, green coverage ratio in villages, and pesticide use per arable land. These indicators reflect essential aspects of the rural environment regarding water supply, drainage, sanitation, and ecological protection, ensuring that residents enjoy a comfortable living space.

Formula: INHAB 2 = drainage pipeline length in villages / rural population; INHAB 5 = pesticide use volume / arable land area.

Indicators of INDOC

The indicators are in Table 4. Improved education opportunities, facilities, and status enhance villagers’ skills and knowledge, fostering local industry development and RR. This study uses five indicators to assess this dimension: cultural center number per capita, percentage of education expenses in total expenditure, proportion of highly qualified teachers in the compulsory education system, full-time teacher number per capita, and rural internet penetration rate. These indicators provide insights into education, indoctrination quality, and modernization in rural areas.

Formula: INDOC 1 = cultural center number in villages / rural population; INDOC 2 = education expenses / total consumption expenditure of rural residents; INDOC 3 = teachers with bachelor and master degrees in middle schools / total teachers in middle schools; INDOC 4 = full-time teachers / rural population; INDOC 5 = internet users / rural population.

Indicators of INFRA

The indicators are in Table 5. Infrastructure provides essential services for living and production, promoting equitable RR. This study uses five indicators to assess infrastructure and governance: road area per capita, electricity consumption per capita, clinic staff per capita, minimum living allowance ratio, and public investment proportion of total annual investment. These indicators reflect the status of roads, capital, electricity, medical care, and social welfare, highlighting government efforts to enhance RR.

Formula: INFRA 1 = road area in villages / rural population; INFRA 2 = electricity consumption in villages / rural population; INFRA 3 = clinic staff number / rural population; INFRA 4 = villagers receiving minimum living allowance / rural population; INFRA 5 = construction investment in public facilities / total annual investment in villages.

Indicators of INCOM

The indicators are listed in Table 6. Income and consumption reflect the material and monetary conditions of rural residents. This study selects five key indicators: urban–rural income gap, rural Engel coefficient, per capita living space, per capita consumption, and per capita living expenditure. These indicators illustrate earning with usage, highlighting the current standard of living for villagers—one of the goals of RR.

The indicator system based on the 5I Framework is comprehensive, representative, and user-friendly (Fig. 2). The color shades of the indicators in Fig. 2 align with those of the connotation in the 5I Framework (Fig. 1), effectively expressing its concepts and assessing RR; for instance, the first three indicators in INDUS in Fig. 2 have the same color shades as the first connotation in INDUS in Fig. 1 (agricultural growth), demonstrating that these indicators represent the first connotation’s meaning. Besides, it balances simplicity with depth without oversimplifying complex issues, affirming that there is an evaluation system with easily understandable indicators for widespread application. In summary, the 5I Framework and its indicator system are memorable and easy to disseminate, contributing theoretically to RR research and highlighting this study’s novelty.

Indicator System Based on the 5I Framework.

Full size image

Methods

After determining the indicators, appropriate methods must be selected. This study employs the entropy weight method for evaluating indicator weights, the technique for order preference by similarity to an ideal solution (TOPSIS) to assess RR’s ODS, fuzzy set qualitative comparative analysis (fsQCA) for analyzing RR paths, and hierarchical clustering to categorize differentiated RR paths. Additionally, this research does not address technology adoption, allowing for a more focused approach to targets and objectives. Specific methods and procedures are detailed below.

Entropy weight method and TOPSIS

When assigning weights to indicators, subjective methods (like subjective and equal weighting) and statistical approaches (such as principal component analysis) are commonly used. However, these methods have drawbacks. (1) Subjective methods rely on expert judgment, introducing personal bias. (2) Equal weighting treats all indicators equally, potentially overlooking more significant ones. (3) Data disturbances can affect statistical methods and may neglect less prominent indicators⁹⁵. To address these issues, this study employs the entropy weight method—an objective approach that assigns weights based on entropy derived from standard data values. That reduces personal subjectivity and considers all indicators for a more reliable evaluation outcome^96,97. Previous studies have widely recognized it for determining indicator weights^74,95. However, it does have limitations; it is sensitive to data quality and initial indicator selection⁷⁴. In this study, values were checked prior to research with no extreme data found, and indicators underwent rigorous selection processes. Thus, the limitations of the entropy weight method can be mitigated.

Various methods like group decision-making and analytic hierarchy processes are used when assessing RR’s overall development status (ODS) across different regions. However, these methods rely on experts’ experiences and opinions, which can be subjective and biased due to limited authority⁸⁴. In contrast, the technique for order preference by similarity to an ideal solution (TOPSIS) offers greater objectivity by ranking alternatives based on their proximity to the ideal solution. That allows for a clearer understanding of relative performances among options⁸⁴. TOPSIS operates on the principle that alternatives should be closest to the best solution while being farthest from the worst. The ideal solution represents a hypothetical maximum or minimum value within the dataset. TOPSIS identifies optimal and worst-case scenarios from a limited set of solutions, calculating distances between evaluation objects and the best scenario to determine their ranks objectively. This ranking provides a clear comparison of alternatives’ relative importance, enhancing objectivity in assessment^85,95. While TOPSIS has limitations—such as assuming linear relationships between indicators and being unable to capture more complex dynamics—it remains effective for exploring ODS dynamics in this study (TOPSIS is effective enough in this study)⁹⁸.

The combination of the entropy weight method and TOPSIS is suitable for this study due to several advantages: (1) They objectively determine the importance of indicators and rank alternatives, enhancing the credibility of assessment results. (2) Their successful application in evaluating various systems ensures more precise and replicable outcomes. (3) They focus on relative performance rather than absolute differences, making comparisons easier^8,99,100 .

Method of fsQCA.

There are various methods for path analysis, including clear set qualitative comparative analysis (csQCA) and multi-value set qualitative comparative analysis (mvQCA). However, they have limitations: (1) csQCA struggles with a large number of interval variables, restricting its use in complex data; (2) mvQCA is not widely applied, leading to underdeveloped theories and tools¹⁰¹. In contrast, fuzzy set qualitative comparative analysis (fsQCA) combines qualitative and quantitative approaches to examine the relationship between antecedent conditions and outcomes¹⁰². It accommodates case complexity and has broader application potential¹⁰¹. Additionally, fsQCA can highlight “differentiated” paths that different villages may take to achieve RR, making it more suitable for this study. While fsQCA has inherent limitations—such as oversimplifying nuanced differences due to its reliance on set membership values¹⁰³- this study prioritizes identifying and clustering main differentiated paths over subtle distinctions; therefore, it is appropriate for this study.

Consistency and coverage are key concepts to clarify before fsQCA analysis. Consistency indicates the relationship between the condition variable and the outcome variable, while coverage measures how well the condition sets explain the outcome¹⁰⁴. The formulas for consistency and coverage can be simplified as follows.

\({X}_{i}\) refers to the membership degree of an element in condition combination \(X\). In contrast, \({Y}_{i}\) refers to that in result \(Y\). If the fuzzy-set value of \(X\) is less than or equal to that of \(Y\) and the consistency value exceeds 0.8, \(X\) is deemed a sufficient condition for \(Y\)¹⁰⁵. If the consistency value surpasses 0.9, \(X\) is considered a necessary condition for \(Y\)¹⁰⁶. After identifying sufficient or necessary conditions, coverage can assess how well combination \(X\) explains result \(Y\); better coverage indicates more substantial explanatory power regarding causality¹⁰⁷.

Method of clustering

When clustering paths, methods like fuzzy clustering, K-means clustering, and Density-Based Spatial Clustering of Applications with Noise (DBSCAN) are commonly used. While effective and innovative, these methods have drawbacks: they can be complex to implement or require predefining the number of categories. Thus, they may not suit this study’s need for a straightforward classification without presets^108,109. In contrast, hierarchical clustering organizes data by progressively merging samples into a tree-like structure. While researchers may find it challenging to choose clustering methods and metrics (which is a limitation), this approach enables them to observe the entire process until all samples form a single class¹⁰⁸. Hierarchical clustering is more suitable for this study for several reasons: (1) it does not require presetting the number of clusters, leading to more objective results; (2) it is easy to calculate and widely applied, aligning with the actual context of this research.

Case and data

This study examines 11 provincial regions in western China (Fig. 3) for several reasons: (1) The rural areas exhibit diverse natural resources and development conditions, leading to significant disparities in RR outcomes. For example, Guizhou faces challenges due to its mountainous terrain, hindering infrastructure and industry development¹¹⁰. In contrast, Yunnan benefits from favorable geography and climate, resulting in rapid advancements in modern agriculture and rural tourism¹¹¹. (2) Compared to other regions in China, the agricultural sector and overall growth of villages in western China are considerably weaker, making RR significant and needed. Despite some progress, much work remains¹¹². Therefore, exploring the current RR status and proposing tailored paths with specific countermeasures that can aid similar areas elsewhere is essential. Note that Tibet is excluded due to a lack of statistical data.

Study Area (source: ArcGIS Desktop 10.6; https://www.esri.com/en-us/home).

Full size image

The other regions of China (especially the eastern ones) were excluded due to their advanced villages, which are less globally representative and do not reflect the actual status of many rural areas. In contrast, western Chinese villages are diverse enough to represent RR experiences and share similarities with numerous villages in other countries¹¹³. That makes the study results more replicable and applicable to various rural contexts.

There are 26 indicators in total, with data sourced from the China Statistical Yearbook, China Rural Statistical Yearbook, China Urban–Rural Construction Statistical Yearbook, Educational Statistics Yearbook of China, China Agricultural Machinery Industry Yearbook, and China Health Statistical Yearbook. The data spans from 2009 to 2021; this study excludes 2022–2024 as it commenced in mid-2023 when statistics were only available up to 2021.

These datasets are sourced from official Chinese government statistics, ensuring high authority, accuracy, and accessibility. To address limited missing data, we use averaging and moving average methods. While these approaches may introduce slight bias or obscure trends, they provide relatively objective estimations compared to more subjective methods (like the Delphi method) and have been widely recognized in previous analyses¹¹⁴. Additionally, to reduce the impact of extreme values and enhance comparability, we calculate “per capita” or “ratio” data for greater accuracy.

Process

ODS Using TOPSIS

(1) Standardize data

The data should be standardized to eliminate unit differences. With 11 regions and 13 years of observations, the total data size for each indicator is 11*13 = 143.

where \(i\) denotes the region, \(j\) denotes the indicator, \({x}_{ij}\) is the original indicator value, and \({x}_{ij}{\prime}\) is the standardized data; Eq. (3) applies to positive indicators, while Eq. (4) applies to negative indicators.

(2) Calculate the entropy value \({e}_{j}\) of the indicator \(j\).

(3) Calculate the coefficient \({g}_{j}\) of the indicator \(j\).

(4) Calculate weights \({W}_{j}.\)

(5) Calculate the positive and negative ideal solution \({X}^{+}\), \({X}^{-}\).

(6) Calculate the positive and negative ideal solution’s distances \({D}_{i}^{+}\), \({D}_{i}^{-},\)

(7) Calculate ODS.

The ODS value typically ranges from 0 to 1. Previous studies often categorize ODS values into five or ten equal intervals (0.2 or 0.1 intervals, respectively)^33,74,115. This paper adopts a ten-category system for more precise differentiation of ODS values, as detailed in Table 7.

Differentiated path using fsQCA

(1) Determine cases, conditions, and outcomes.

This paper employs fsQCA using data from 2019 to 2021 to explore the differentiated path of RR within the 5I Framework. The software used is fsQCA 3.0. The analysis includes five antecedent conditions (the five “I” dimensions) and 11 cases (regions in western China), with the outcome of ODS meeting medium case selection requirements¹¹⁶.

(2) Calibrate data.

Before conducting the necessary condition analysis, it is essential to calibrate the data for conditions and results. Data calibration assigns membership degrees to cases through standard methods: indirect and direct calibration. Indirect calibration adjusts measurements based on qualitative assessments, while direct calibration is preferred for its rigor and reliability^117,118. Three anchor points are established in direct calibration: full membership point, crossover point, and full non-membership point. This process converts conditions and results into a fuzzy set with values ranging from 0 to 1, indicating case membership degree. This study utilizes a direct calibration method with anchor points at corresponding fuzzy set values of 75%, 50%, and 25% as full membership points, crossover points, and full non-membership points, respectively—widely adopted in previous research for accuracy^119,120. Results are presented in Table 8.

After calibration, cases with fuzzy set values of 0.5 should be excluded¹¹⁶. For data validity, the value can be adjusted from 0.5 to 0.501, a common practice in analysis¹⁰⁶.

(3) Analyze necessary conditions.

Evaluating whether each dimension (namely condition) of the 5I Framework is necessary for achieving high and low RR results is essential. It is deemed necessary if condition “I” consistently appears when high or low RR occurs. Necessary condition analysis assesses the membership degree of set “I” within the output set of ODS. Coverage is significant only for conditions that pass the consistency test (i.e., necessary conditions). Typically, coverage for these conditions should exceed 0.5 to measure their empirical correlation effectively¹²¹. If an “I” condition proves necessary, it must be retained in subsequent sufficiency analyses (i.e., differentiated paths analysis).

Table 9 presents the results of the necessary condition analysis for high and low ODS of RR. The consistency for each condition falls below the threshold of 0.9¹²¹, indicating that multiple dimensions of the 5I Framework may influence ODS outcomes in Western China.

(4) Construct the truth table.

The truth table outlines condition combinations that lead to outcomes, which can be simplified using frequency and consistency thresholds. The frequency threshold should cover over 75% of cases¹¹⁸. Additionally, the consistency threshold must be at least 0.75¹¹⁸.

In this study, the frequency threshold is set to 1 and the consistency threshold to 0.8, as suggested in other studies^119,121. Furthermore, PRI consistency (proportional reduction in inconsistency) is evaluated as an alternative method for assessing the consistency of subset relations while avoiding simultaneous subset relations. The PRI threshold value is established at 0.75, a common standard in previous studies^106,122.

(5) Sufficiency analysis (differentiated paths of RR).

The 5I dimensions influence RR in western regions through various combinations, necessitating a sufficiency analysis to identify the distinct paths that lead to high and low ODS outcomes. Using fsQCA, three types of solutions are derived: complex, parsimonious, and intermediate solutions. “Solutions” refer to path combinations supported by cases¹²². Complex solutions enumerate all possible combinations but can be challenging to interpret due to their abundance. These can be simplified into parsimonious and intermediate solutions; the former is more straightforward as they highlight critical non-negligible conditions based on simplifying assumptions. Intermediate solutions arise from counterfactual analyses of complex and parsimonious solutions, including only theoretically plausible counterfactuals.

Previous research indicates that core conditions appear in parsimonious and intermediate solutions, while conditions unique to intermediate solutions are deemed peripheral^91,106. This study’s core conditions signify a strong causal link between the “I” dimension and ODS outcomes, while peripheral conditions indicate weak causality.

(6) Test the robustness.

In this paper, the consistency and PRI thresholds are used in the robustness test for sufficiency analysis. The consistency threshold is adjusted from 0.8 to 0.75, and the PRI threshold is adjusted from 0.75 to 0.8 to identify paths leading to high ODS⁹¹.

Cluster classification

(1) Convert conditions to numerical data sets.

Hierarchical clustering requires numerical data for calculations¹²³. This study converts conditions into numbers with equal intervals (Table 10). The specific values are unimportant as long as the intervals are consistent. Additionally, high ODS and low ODS have opposite meanings and, thus, opposite values.

(2) Use SPSS 20.0 for hierarchical cluster analysis.

This study does not specify the number of clusters in advance; the “between groups” method and “Euclidean distance” are used as the metric. Data are not standardized, and “Variable” is selected in the “Cluster” box, a common approach in previous studies¹⁰⁸. The dendrogram is generated using average linkage.

Descriptive statistics

The descriptive statistics of the indicators and regions aim to help readers understand the case’s features and facilitate comparisons with other contexts. Table 11 presents the indicators’ average values, standard deviations, and weights (11 regions over 13 years). Notably, weights are derived using the entropy weight method, with a data size 143 for each indicator (11 regions * 13 years). Precisely, INDUS 2 (agricultural labor productivity rate) and INHAB 4 (green coverage ratio in villages) have the highest weights (0.146 and 0.144), indicating significant data variance among regions in these areas as well as their importance in RR.

Table 12 shows the 11 selected regions’ rural statistical characteristics for 2022. The data, sourced from the China Statistical Yearbook, ensures authority. Notable disparities exist; for example, Sichuan has a population 15 times that of Ningxia, while Inner Mongolia’s per capita disposable income is 1.6 times higher than Gansu’s. These differences highlight the inequalities in rural resources and RR outcomes, underscoring the need to explore their distinct paths to RR and demonstrating their representativeness of similar villages.

Analysis of ODS

The descriptive statistics highlight disparities among villages in these regions, necessitating an exploration of their RR status to identify potential paths for achieving RR. The following sub-section addresses the research question 3–1: “What is the current status and differences of RR?” Possible reasons are analyzed through the 5I Framework, facilitating easier understanding for readers.

Temporal trends

ODS primarily assesses RR performance across regions. Figure 4 illustrates the temporal variations of ODS, which generally range from 0.3 to 0.6, indicating three grades: acceptable, moderate, and good. The fluctuations within neighboring grades suggest that RR has remained relatively stable in recent years. However, none of the recent ODS values have reached “very good” status (ODS ≥ 0.6), highlighting the potential for improvement in RR. Moreover, the narrowing gap in ODS indicates that RR positively contributes to reducing regional rural disparities. Notably, many regions have experienced a slight decrease in ODS since 2020, likely due to the impacts of the COVID-19 pandemic; this suggests that unforeseen events may influence RR efforts endogenously.

Temporal Trends of ODS.

Full size image

Regional comparison in time

(1) The ODS of villages in Ningxia and Inner Mongolia has consistently been better (Grade Good) than others, but their trends differ; Ningxia is declining while Inner Mongolia is improving. Specifically, Inner Mongolia has recently experienced rapid growth in the proportion of agriculture-added value in the gross regional product and grain output per capita, enhancing the INDUS dimension of the 5I Framework. In contrast, Ningxia’s former indicator has declined. This decline does not indicate a deterioration in agriculture (INDUS); instead, it reflects a more significant emphasis on agriculture in Inner Mongolia and other non-agricultural industries in Ningxia.

(2) Chongqing’s villages show a downward trend in ODS from 0.439 in 2009 to 0.407 in 2021 for two main reasons. Firstly, as a rapidly developing municipality, Chongqing faces swift urbanization that diminishes village attractiveness (INHAB decrease). Urban-related industries like tourism and advanced manufacturing have absorbed significant employment opportunities, leading to rural labor outflow that should have supported local RR processes. Secondly, limited land for specialized agriculture results in low yields and hinders high-quality agricultural growth (INDUS decrease).

(3) The ODS in rural areas of Guangxi and Yunnan has shown minimal fluctuations, consistently remaining at a lower grade (Acceptable). That can be attributed to several factors. Firstly, remote locations and poor transportation hinder better ODS of RR; many villages are hard to access, and high logistics costs limit their potential for improvement (low INFRA). Secondly, rural infrastructure requires enhancement. Limited local financial support has slowed the development of essential facilities like roads, electricity, and communications, which impedes ODS (low INFRA). Lastly, there is a lack of educational resources in these areas. Students in remote villages do not have access to quality education, affecting overall indoctrination quality (low INDOC).

Spatial differences

Figure 5 illustrates the spatial variations of ODS in 2009, 2012, 2015, 2018, and 2021. Overall, rural areas in western China show distinct spatial differences in ODS. The specific results are as follows.

Spatial Variations of ODS (source: ArcGIS Desktop 10.6; https://www.esri.com/en-us/home).

Full size image

(1) The ODS of villages in northern regions is generally higher than that of southern and central areas. Villages in Inner Mongolia, Ningxia, Shaanxi, and Chongqing exhibit relatively high ODS in Moderate and Good grades. In contrast, Qinghai, Gansu, Yunnan, and Guangxi are more stable but lag. Geographic limitations significantly affect RR in these backward regions (low INDUS). For instance, Gansu suffers from severe soil erosion (11,505.1 thousand hectares affected in 2022) despite having a smaller land area than Sichuan; Qinghai’s ___location on the high-altitude Tibetan plateau poses challenges; Guangxi’s hilly terrain (14.49% hills) and Yunnan’s mountainous landscape (88.6% mountainous area) hinder agricultural expansion^124,125. These natural constraints complicate sustainable rural development. In contrast, regions with higher ODS benefit from various advantages: Inner Mongolia has a developed mining industry providing villagers with substantial incomes while local governments can subsidize farmers (high INCOM); Shaanxi’s Guanzhong Plain offers abundant high-quality arable land conducive to agriculture alongside effective environmental governance on the Loess Plateau (high INDUS and INHAB); Chongqing experiences frequent urban–rural resource flow allowing rural areas to leverage city spillover effects for better public education and infrastructure investment (high INDOC and INFRA). These factors contribute to the elevated ODS of RR in these regions.

(2) Xinjiang’s rural ODS is notably more substantial than the surrounding regions. Despite its remoteness, Xinjiang’s villages benefit from several comparative advantages: first, it has a solid agricultural foundation with high-quality arable land yielding profitable crops like cotton, wolfberry, and sugarbeet; second, a specialized agricultural reclamation group employs innovative production models such as “digital and precise agriculture” to enhance output; third, frequent agricultural trade with Central Asian countries fosters growth; fourth, national authorities prioritize border and minority issues in Xinjiang, providing substantial annual support for village development (resulting in relatively high INDUS and INFRA).

(3) ODS shows apparent fluctuations and synchronicity in Guizhou and Sichuan. Low ODS levels were initially observed in both areas (the 2008 earthquake notably impacted Sichuan). However, government support led to continuous agricultural growth alongside investments in rural infrastructure and housing (increased INFRA and INHAB). By 2012, both regions’ villages achieved improved ODS grades due to the rise of rural tourism. Recently, COVID-19 disrupted this progress, causing a decline back to lower grades by 2021.

RR’s ODS temporal and spatial analysis is dynamic and holistic while allowing for comparison. That enhances readers’ understanding of current developmental dynamics among villages—a novel aspect of this study. The analysis effectively addresses research question 3–1 regarding the current status of RR and its differences while providing references for further exploration.

Differentiated paths of RR

Based on ODS status, this sub-section explores differentiated paths to achieve RR, addressing the question 3–1: “How can various villages enhance aspects to achieve RR?”.

Table 13 presents the fsQCA analysis results, highlighting paths leading to high and low ODS of RR within the 5I Framework’s dimensions. Paths 1–4 yield high ODS, while Paths 5–7 result in low ODS. The overall coverage for solutions (differentiated paths) resulting in high RR is 0.666, with a consistency of 0.941—exceeding the acceptable threshold of 0.75¹²⁶. That indicates that these four paths account for 66.6% of case regions and effectively illustrate diverse routes to high ODS. For solutions causing low ODS, coverage is at 0.662 with an overall consistency of 0.975, representing 66.2% of total case regions; thus, the findings are valid. The symbol “*” denotes logical “AND,” while “ ~ ” signifies the absence of a condition.

Differentiated paths to high ODS

(1) Path 1: INDUS* ~ INHAB*INDOC*INFRA* ~ INCOM. The consistency of Path 1 is 0.820, with a raw coverage of 0.189. In this path, INDUS and INFRA are core conditions essential for high ODS, while INDOC is a peripheral condition that contributes significantly to it. INHAB and INCOM are absent peripheral conditions that play an unimportant role. ODS requires support from the agricultural industry, robust infrastructure, and educational development to achieve RR. Under these conditions, villages can transcend the local living environment and income constraints to achieve RR.

The representative case for Path 1 is Xinjiang. Regarding INDUS, Xinjiang leverages its vast territory for large-scale agriculture and animal husbandry, achieving industrialized production and exporting products to Central and West Asia due to its geographical advantages. For INFRA, the local government prioritizes infrastructure development; improved highways enhance connectivity with neighboring areas for better logistics and trade. Regarding INDOC, the government invests in rural education by initiating exchange programs with developed regions to boost villagers’ educational status and cultural literacy—especially among ethnic minorities.

(2) Path 2: INDUS* ~ INHAB* ~ INDOC*INFRA*INCOM. The consistency is 1, and the raw coverage is 0.260, the highest among high ODS paths. INDUS, INFRA, and INCOM are core conditions; INHAB and INDOC are absent peripheral conditions. This path indicates that high ODS in RR relies on advantages from agriculture, infrastructure, and improved living standards.

Notable cases include Inner Mongolia and Chongqing. In INDUS, Inner Mongolia leverages its grassland and natural pastures as a key livestock base in China. Meanwhile, Chongqing’s mountainous villages utilize abundant water resources to grow cash crops like vegetables, fruits, and tea. In INFRA, Inner Mongolia has significantly invested in rural construction and road upgrades to enhance agricultural product transportation efficiency; Chongqing has similarly improved rural roads and clinics to support RR. In INCOM, Inner Mongolia focuses on boosting villagers’ income through subsidies, public insurance, and low-interest loans; Chongqing targets poverty alleviation with tailored strategies across industry, education, and healthcare sectors.

(3) Path 3: ~ INDUS*INHAB* ~ INDOC*INFRA*INCOM. This path has a consistency of 1 and a raw coverage of 0.208. INFRA and INCOM are core conditions, while INHAB is peripheral. Unlike the previous paths that depend on improved agricultural industries, this one requires good infrastructure, stable income and consumption for residents, and a suitable living environment. With these conditions, villages can transcend the limitations of agriculture and rural education (INDUS and INDOC) to achieve sustainable RR.

A representative case is Shaanxi. Regarding INHAB, the local government initiated the rural housing renovation program and increased investment in ecological governance while implementing policies like dangerous house renovations, migrant villager relocations, and eco-friendly countryside construction. For INFRA, villages have launched power grid upgrades to address voltage quality issues and ensure electricity supply. Regarding INCOM, villagers earn income through non-agricultural work as migrant workers in cities; additionally, the local government offers legal protection and skill training to secure their legitimate earnings.

(4) Path 4: INDUS*INHAB*INDOC*INFRA*INCOM. This path has a consistency of 0.985 and raw coverage of 0.497. INDUS, INFRA, and INCOM are core conditions, while INHAB and INCOM are peripheral conditions. It indicates that exceptional ODS requires support across all dimensions.

The representative case is Ningxia, where development across five dimensions enhances RR’s ODS. In INDUS, Ningxia’s Yellow River irrigated agricultural area is China’s key grain production zone. Villages utilize water-saving irrigation and optimize planting structures to cultivate efficient crops like wolfberry and hair weeds, advancing deep processing of agricultural products and agritourism. In INHAB, ecological protection projects—such as sand control, converting farmland to grassland, and soil conservation—have made villages more environmentally appealing. In INDOC, Ningxia was among the first regions in China to launch an “internet + education” program; village students benefit from shared educational resources and innovative teaching tools for improved rural education. In INFRA, Ningxia has upgraded rural roads, enhanced transportation networks, renovated power grids, and expanded minimum living allowance services, ensuring villagers live with dignity. In INCOM, the local government offers targeted employment subsidies for women and the elderly while attracting leading enterprises and expanding collective economy initiatives to improve villagers’ quality of life.

Differentiated paths to low ODS

The analysis of differentiated paths contributing to low ODS explores the challenges in achieving RR without specific dimensions. The paths leading to low ODS differ significantly from those resulting in high ODS, as illustrated in Paths 5–7.

• (1) Path 5: ~ INHAB*INDOC* ~ INFRA* ~ INCOM. Here, INDOC represents the current peripheral condition, while INHAB, INFRA, and INCOM are absent peripheral conditions. That suggests that even with ideal educational support, better rural revitalization is only possible with a favorable living environment, government backing, and guaranteed incomes for residents—evident in Guangxi.

• (2) Path 6: ~ INDUS*INHAB*INDOC* ~ INCOM. In this case, INHAB is the present peripheral condition; INDOC is the present core condition, while INDUS and INCOM are absent core conditions. It indicates that rural areas struggle to achieve high ODS without proper agricultural growth and stable villagers’ income despite having suitable living environments and education—seen in Guizhou and Yunnan.

• (3) Path 7: INDUS*INHAB* ~ INDOC* ~ INFRA*INCOM. Here, INDUS is a present core condition; INHAB and INCOM are present peripheral conditions, while both INDOC (absent peripheral condition) and INFRA (absent core condition) hinder progress. This path shows that achieving RR is challenging without sufficient indoctrination or infrastructure development—even if residents enjoy improved livelihoods, good living conditions, and prosperous agriculture.

Overall, INFRA is a core condition in all high ODS paths, highlighting its essential role in promoting RR. Government investment in rural infrastructure ensures energy supply, transportation, health, and welfare services—key elements for sustaining high ODS. Additionally, INDUS and INCOM are significant conditions for advancing RR; INDUS is vital in Paths 1, 2, and 4, while INCOM is critical in Paths 2, 3, and 4, proving that enhancing agricultural value and increasing villagers’ income and consumption levels greatly accelerate RR. In contrast, INHAB and INDOC are not at the core of every path.

Furthermore, robustness tests confirm that neither method produces different path combinations or shows significant changes in number, consistency, or coverage. Thus, the results pass the robustness test convincingly addressing research question 3–2: “How can various villages enhance aspects to achieve RR?”.

To help readers quickly summarize these specific paths, Fig. 6 visualizes the paths to achieve high and low ODS of RR based on the 5I Framework, highlighting key dimensions in each path. The colors of the dimensions correspond with those in the 5I Framework (Fig. 1) to clarify its application scenarios. Present dimensions (core or peripheral conditions) are shown; absent or non-essential ones are omitted. Notably, paths 5–7 indicate routes leading to low ODS, allowing readers to focus on missing dimensions—suggesting that “without these dimensions, high ODS for RR cannot be achieved even if other dimensions exist.”

Visualization of the Differentiated Paths.

Full size image

Categories of differentiated paths

For proposing effective and replicable countermeasures for enhancing the ODS of RR, seven differentiated paths were clustered using hierarchical clustering to address research question 4–1: “Are there commonalities in the various paths to implementing RR?”.

Four distinct categories emerged based on their characteristics. The first category emphasizes that improving all dimensions contributes to RR; the second focuses on education (villagers’ learning); the third highlights labor devotion; and the fourth addresses the living environment. These features can be summarized as Lots, Learning, Labor, and Living—initially represented by "L." Thus, these differentiated path clusters are termed the “4L Category.” Details of the 4L Category are illustrated in Fig. 7 below.

4L Category.

Full size image

Category 1: lots

This category includes Path 4, highlighting that RR can be achieved by integrating “lots” of aspects such as agricultural development, living environment enhancement, rural education, infrastructure improvement, and residents’ income. Ningxia exemplifies this success with notable ODS performance across all these aspects. In this category, diverse stakeholders—local governments, enterprises, village communities, and villagers—can collaboratively enhance all dimensions of RR.

Category 2: learning

This category includes Paths 1 and 7, accounting for 31.3% of cases. Unlike Category 1, it suggests that villages may leverage specific dimensions while temporarily neglecting others to achieve differentiated RR. Category 2 highlights that RR in some villages largely depends on enhancing “learning” and cultural status among residents. Strategically, rural culture and education can be advanced through two key approaches: optimizing educational resource allocation and integrating digital technology with rural education.

Category 3: labor

This category includes Paths 2 and 6, accounting for 53.4% of cases. It indicates that improving RR requires villagers’ “labor” to support the agricultural industry. While other aspects may exist within this category, agricultural development is crucial, necessitating collective effort from villagers. Strategically, rural areas can boost agricultural efficiency by optimizing management practices and enhancing crops’ disaster-prevention capabilities.

Category 4: living

This category includes Paths 3 and 5, accounting for 70.5% of cases, the highest proportion. It suggests that rural areas can achieve RR by enhancing the “living” environment, which supports other aspects of development. Strategic improvements to the living environment can be made from ecological, management, and living experience perspectives.

The 4L Category highlights commonalities in the differentiated paths of RR, making it easy to apply globally and effective for targeted countermeasures. It addresses the research question 4–1: “Are there commonalities in the various paths to implementing RR?”.

Notably, the “5I Framework” is a theoretical evaluation tool analyzed from five aspects, while the “4L Category” combines various RR paths derived from this framework. Essentially, the “5I Framework” underpins the “4L Category,” representing its practical application. They are strongly correlated both theoretically and practically.

Countermeasures

Different villages follow distinct paths of RR. According to the “4L Category,” villages in various categories require tailored countermeasures to achieve their specific RR paths. This subsection addresses the research question 4–2: “What specific actions can villages in different contexts take to achieve RR?” These countermeasures also apply to similar villages in other regions or countries.

Countermeasures for category 1 (Lots)

For Category 1 (Lots), it is essential to coordinate and harmonize all dimensions to achieve comprehensive growth of RR. The countermeasures are as follows.

• (1) Promote efficient agricultural technologies. Local governments should support adopting advanced agricultural practices and encourage recognized machinery companies to participate in technology upgrades through subsidies.

• (2) Enhance villagers’ living conditions. Local governments can invest in building high-standard, modernized, rural-collectively-owned apartments to ensure residents’ living conditions. They can also invite professionals to direct agricultural production activities to reduce pollution.

• (3) Improve rural education. Increased investment in vocational training, practical courses in agriculture management, upgraded online resources, and faster internet access will enable effective education for villagers.

• (4) Attract investments for infrastructure improvement. Implement favorable policies to attract foreign and domestic investment in infrastructure upgradation. Financial investments can enhance road conditions for better communication, and intellectual investments can boost clinic performance to increase villager satisfaction.

• (5) Diversify income opportunities by offering suitable employment options. For instance, some jobs can be shared by babysitters or part-time workers, job positions in surrounding towns can be provided for those who accept commuting, and specially provided jobs in governments or organizations can be offered to disabled individuals.

Countermeasures for category 2 (Learning)

For Category 2 (Learning), promoting rural education and enhancing villagers’ literacy is essential. The countermeasures are as follows.

(1) Improve the quality of rural teachers. That can be achieved through state-funded training programs, teacher rotation, academic visits, resource sharing between urban and rural areas, and increased salaries.

(2) Optimize educational facilities in rural schools. Leveraging corporate sponsorships, charitable donations, and volunteer services can ensure a conducive learning environment for all children.

(3) Integrate digital technology with education to achieve a “digit + education” model. Schools can utilize national online platforms and international databases to enrich teaching resources, particularly in agriculture and ecological conservation.

(4) Implement vocational training programs tailored to local needs. Stakeholders should offer courses based on regional characteristics and industrial demands. Courses aiming to enhance villagers’ skills and industry performances will be welcomed. Topics like modern agricultural techniques, product processing, e-commerce, and rural tourism will be especially relevant.

Countermeasures for category 3 (Labor)

For Category 3 (Labor), optimizing local agricultural operations is essential. The countermeasures are as follows:

(1) Foster partnerships among enterprises, higher education institutions, and research organizations to drive innovation in the agricultural industry. Enterprises can invest in university research to explore advanced techniques like water-saving irrigation, improved seed breeding, and biological pest management; scholars can offer guidance on agricultural production and management.

(2) Develop integrated digital platforms to streamline agricultural resources and production. These platforms can coordinate digital machines (e.g., drones) for precision farming, facilitate efficient processing of agricultural products, and establish business collaborations.

(3) Offer attractive financial services tailored for agriculture that lower financing costs for enterprises and villagers. To obtain better financial services, villages can negotiate with service providers by offering incentives such as waiving rent for businesses, providing free meals to staff, and promising minimum annual profits.

(4) Improve the quality of the local labor force. That can be achieved through agriculture-related training programs that equip villagers with new skills and ideas. Enhancing work environments will also attract and retain managerial talent, further elevating labor quality.

Countermeasures for category 4 (Living)

For Category 4 (Living), ensuring villagers’ primary living conditions is essential. The countermeasures are as follows.

(1) Ensure water safety. Establish water purification plants for clean drinking water, build sewage treatment facilities to safely discharge agricultural and domestic waste, and impose penalties on individuals or enterprises that pollute water sources.

(2) Improve sanitation and housing standards. New houses must include toilets with uniform sewage treatment; renovated homes meeting modern standards (such as high safety and environmental sustainability) can receive subsidies; high-standard garbage stations should be established for sorting and recycling waste.

(3) Promote cleaner energy. Encourage using clean energy sources (like solar and bioenergy) and support villagers in building distributed energy systems to reduce environmental pollution while ensuring adequate rural energy supplies.

(4) Create an environmental governance mechanism with villager participation. Develop inclusive policies that promote self-governance, encourage community involvement in decision-making regarding environmental issues, and raise awareness about environmental protection among villagers.

Countermeasures for Other regions to achieve sustainable rural development

These countermeasures also reference regions or countries wishing to achieve sustainable rural development.

(1) Agriculture is essential: Optimize agricultural structure, develop modern technologies, and promote agricultural industrialization to enhance production efficiency and product quality.

(2) Ecology is essential: Consider the living conditions of villagers and the ecological environment; rural development must not come at the cost of ecological destruction.

(3) Cultural-ethical development is essential: Villagers need material and spiritual prosperity. Integrate traditional culture, learning resources, and digital technology to enrich their spirits anytime, anywhere.

(4) Infrastructure is essential: Improve rural roads, water supply, electricity, internet access, schools, clinics, and entertainment facilities to enhance villagers’ lives.

(5) Multi-body participation is meaningful: National authorities, local governments, village communities, villagers themselves, enterprises, and universities are all stakeholders in sustainable rural development. Their coordinated efforts contribute to better growth.

The “4L Category” countermeasures illustrate differentiated paths to RR, offering practical insights and novelty that address the research question 4–2: “What specific actions can villages in different contexts take to achieve RR,” thus accelerating future practices in RR.

Core results

Differentiated paths of RR are crucial for sustainable rural development. This study analyzes 11 regions in western China using TOPSIS and fsQCA methods, constructing a 5I Framework with 26 key indicators to assess RR. The research evaluates the ODS of these regions over the past decade, explores differentiated paths of RR, categorizes them into four groups (“4L Category”), and proposes targeted countermeasures. The raw data is publicly available, the indicators are derived from explicit formulas, the calculation processes are objective and detailed, and the thresholds are widely accepted; they enhance the objectivity and replicability of the results. Key findings include:

1. (1)

   The 5I Framework—comprising INDUS (industry), INDOC (indoctrination), INHAB (inhabited environment), INFRA (infrastructure), and INCOM (income)—effectively reflects agricultural growth, village development and residents’ expenditure through its 26 indicators. This framework simplifies the understanding of RR globally. It provides an accessible evaluation system, addressing research questions 1 and 2: “Is there a clear and explicit RR framework that simplifies the global theoretical understanding of RR?” and “Is there an evaluation system with easily understandable indicators for widespread application?”.

2. (2)

   The ODS of RR in western China shows stable fluctuations at adjacent grades with significant room for improvement; in detail, ODS fluctuates between the Acceptable and Good Grades (0.3 and 0.6) with Chongqing’s villages showing a downward trend (from 0.439 to 0.407). Villages in northern regions generally outperform southern and central areas. Xinjiang exhibits notably higher ODS (Grade Good) than neighboring regions, while Guizhou and Sichuan show marked fluctuations in village-level ODS. These results address the research question 3–1: “What is the current status of RR and its differences?”.

3. (3)

   Seven differentiated paths to ODS of RR were identified, indicating that different villages have unique approaches. INFRA is essential in all high ODS paths, while INHAB and INDOC are not always central. That addresses the question 3–2: “How can various villages enhance aspects to achieve RR?”.

4. (4)

   The seven paths are grouped into four replicable categories—4L Category: Lots, Learning, Labor, and Living – emphasizing that RR can be achieved by integrating “lots” of aspects, enhancing “learning” and cultural status among residents, requiring villagers’ “labor” to support the agricultural industry, or enhancing the “living” environment, and highlighting distinct targets and strategies for each village. That answers research question 4–1: “Are there commonalities in the various paths to implementing RR?”.

5. (5)

   Specific countermeasures for implementing RR within the “4L Category” are necessary. For Lots, coordination across dimensions is vital for comprehensive growth; for Learning, promoting rural education and improving literacy is crucial; for Labor, optimizing local agricultural operations is needed; and for Living, ensuring basic living conditions for villagers is essential. That responds to research question 4–2: “What specific actions can villages in different contexts take to achieve RR?”.

Innovations

This study presents the following innovations.

1. (1)

   A new 5I Framework with an indicator system is proposed, addressing multidimensional factors influencing RR. This framework simplifies complex RR issues through precise components, making it easy to understand and disseminate; it contributes theoretically to RR research, providing new insights into RR to the existing literature.

2. (2)

   The combination of the entropy weight method and TOPSIS objectively evaluates the relative overall development status of RR, and fsQCA explores the differentiated paths to RR; they provide objective, replicable, and useful methodological references for other studies in other contexts.

3. (3)

   The overall development status of RR is analyzed both temporally and spatially. This dynamic, holistic, and comparative perspective enhances readers’ understanding of development dynamics and village differences.

4. (4)

   Differentiated paths to RR integrated into the “4L Category” emphasize specific countermeasures that villages adopt in various contexts for effective RR; this integration supports practical applications in RR efforts.

Limitations and future research directions

Firstly, this study focuses on villages in western China, and findings may not apply to villages with different socio-economic, cultural, or geographic characteristics elsewhere; thus, caution is needed in interpretation. Secondly, some indicators and data may be unavailable in specific regions and villages due to the publication timeliness and completeness of official statistical yearbooks; that leads to potential gaps in critical indicators and less precise data than some readers might expect. Thirdly, some indicators may lack relevance outside this context and could be insufficiently detailed or representative.

Future research will explore a broader range of cases and may expand the 5I Framework and 4L Category to include more diverse settings to identify additional differentiated paths for RR. Additionally, future studies could utilize a broader database, enhance geographic coverage, and integrate qualitative analyses to incorporate more representative indicators from a micro perspective (such as village data) alongside quantitative findings; these efforts will improve the precision and applicability of the results.