2.1 Digital village construction

Some countries and regions have already started digital village construction activities. The Information Network Village Project was started in Korea in 2001 with the intention of bridging the enormous urban-rural digital divide and the rural exodus [9]. In 2014, the Chinese city of Beijing launched a pilot digital village construction project to increase the effectiveness and quality of rural industries, as well as the refinement of rural governance and public services. By the end of 2017, Beijing has basically completed the construction of 135 smart villages, covering 13 districts in the city [10]. In 2015, Hainan Province launched the construction of "Internet Agricultural Towns" with the town as the unit. The "1 + 2 + N" operation model was built with "Internet" as its characteristic [11]. In an effort to transition from a conventional society to one that is digitally empowered and knowledge-based, India began the Digital India Programme in 2015 [12]. In 2016, Japan proposed "Next generation agricultural, forestry and water industry Technology based on Intelligent machinery + Intelligent IT project" to apply digital technology to natural resource management and disaster prevention [13]. The EU introduced the Smart Villages campaign in 2017 and began a two-and-a-half-year Smart Villages 21 project in 2019 to pick 21 countrysides to provide monetary and technical help and ultimately scale up experience of building smart villages in 21 EU villages [14]. Under the framework of the Smart Village movement, Italy, France, Poland, and other countries have launched different digital village construction plans according to their conditions.

At present, the theoretical research and practical application of digital countryside are primarily concentrated in developed countries and regions, such as the United States, the European Union [15], Australia [16], Ireland [17], and Poland [18], while little attention is paid to the rural digitalization policies or planning plans in developing countries [19]. At the same time, it can be seen that the core of national digital rural establishment is not identical. India focuses on digital technology infrastructure construction and improvement. European nations primarily improve the environment for rural development by applying new technologies and strengthening public services [20]. The construction of China’s digital villages is more difficult due to environmental differences between China and other nations, as well as the striking differences between regions within China. However, there has been little in-depth exploration of the content and establishment logic of digital villages as well as no in-depth analysis of the practical cases of China’s digital villages.

2.2 Digital village development and evaluation

In the development model of the digital countryside, numerous scholars have conducted research. Arabatzis et al. used the PROMETHEE-II method to evaluate local group actions in implementing LEADER+CI in Greece [21]. Ella et al. used a qualitative research approach to construct a development model for Indonesia that can develop smart villages through an ecosystem approach with a bottom-up planning process as the primary method and a collaborative governance model as the core model [22]. Manasijević et al. construct a development model for the digital transformation of villages that contributes to the relativization of regional disparities in the Republic of Serbia by investigating the advance of digital transformation of Vrmdža village with the Internet and modern digital technologies for engaging in business activities in villages as the core model [23]. Murty et al. take an endogenous approach to the construction of a new development framework for smart villages and divide it into four stages: discovery, planning, outsourcing, and implementation, laying the groundwork for the development of a scalable smart village framework [24]. Kyriakopoulos et al. conducted a comprehensive analysis of the impact of climate change on Greek agriculture through the new agricultural policy, and proposed corresponding critical interventions and actions [25].

Meanwhile, in terms of assessment index system construction, Shen et al. constructed an assessment index system from five aspects from the perspective of agricultural informatization and used the hierarchical analysis method (AHP) to evaluate the degree of agricultural digital development [26]. Based on the connotation of the concept of the digital countryside, Chang et al. constructed an assessment index system for the digital countryside, which contains nine first-level indicators of capability and effectiveness categories, as a way to scientifically evaluate the advanced level of the digital village [27]. Cui et al. explored the theory and designed the main indicators of the rural digital economy around the digital economy composition of industrial digitization and digital industrialization [28]. Zhang et al. measured and evaluated the readiness of China’s provincial digital countryside development by constructing a digital countryside’s advance readiness index system using the AHP method and entropy value method[29]. Mu et al. created an index system of the digital economy in agriculture and rural areas from the industrial level on the basis of three components of digital economy formation [30]. Li Weiwei et al. combined the Boston Consulting Group Matrix and GeoDetector to study the construction level and evolution pattern of digital villages in 70 counties in Guangxi [31]. Zhang Ping et al. used geographic information systems (GIS) to evaluate the development of the digital countryside in Gansu Province and then formulated targeted development strategies according to the development level of each region [32]. Yang Xiaojuan et al. used the AHP and BCG matrix to evaluate the development of rural revitalization in 131 cities in western China and studied rural revitalisation performance’s evolution model and driving mechanism [33].

To sum up, the study of digital countryside has attracted more and more attention. It has made good progress in various fields, providing valuable experience for this paper’s study. However, there are some limitations to the existing research. Firstly, scholars mainly study digital countryside at the theoretical level, focusing on the digital countryside’s connotation, operation mechanism and governance strategy. However, little attention has been paid to the ongoing development of digital villages around the country, and there are no accurate and consistent measurement tools to measure the current development of digital villages correctly. Secondly, most existing studies construct the index system from a specific perspective, such as agricultural informatization, digital economy and industry level, and the coverage of digital rural indicators is insufficient. Factors like farmers’ life services and the rural ecological environment are often ignored. Finally, in terms of evaluation methods, single weighting methods such as analytic hierarchy process (AHP) or entropy value method are generally used. There are biases in subjective judgment or quality biases in objective data, leading to unsatisfactory evaluation results. Therefore, this paper comprehensively considers all aspects of building a sound evaluation index system, including infrastructure, service, economy, life and digital green production. Taking Jiangxi Province as a case, the paper comprehensively evaluates the development of the digital countryside from the provincial level. Then, the CRITIC-G1 weighting method is adopted, which reflects the internal relationship between indicators and the objective characteristics of data. Finally, according to the evaluation results, the development of digital villages in Jiangxi Province is studied and analyzed, and targeted measures are formulated to promote the construction of digital villages. This paper hopes to establish a scientific and reasonable evaluation model to truthfully evaluate the development level of digital villages in various regions and then solve and promote the construction of digital villages according to the reflected problems.

4.1 Improved CRITIC correction G1

At present, many mature methods have emerged for determining the weights of various indicators, such as the AHP method and Delphi method, which rely on experts’ subjective experience, and the weights obtained by these methods are affected by experts’ experience and professionalism. In contrast, objective weighting methods such as the entropy value method, rough set method and standard deviation method that rely on mining objective data information can only reflect objective information and cannot incorporate decision makers’ subjective consciousness into decision-making, which is easily detached from reality. Furthermore, the factors that influence the growth of digital villages are complex and difficult to quantify [29]. Purely subjective or objective empowerment will have an impact on the subsequent comprehensive evaluation, while the comprehensive assignment method combines the benefits of subjective and objective assignment methods, which can reflect the personal will of experts and apply objective theories and methods in the meantime. Therefore, this article selects the comprehensive assignment method that combines subjective and objective.

The G1 weighting method is a modified method of the AHP method, in which experts assign values to ranked indicators based on their importance for a two-by-two comparison and then apply the importance formula to calculate indicator weights [42]. The G1 method can fully reflect the subjective consciousness of the evaluator and find out the intrinsic connection of the importance of the influencing factors based on the sequential relationship between the influencing factors, without testing for consistency. It both ensures the authenticity of subjective judgment and enhances the simplicity of operation. However, this weighting method, which relies on experts’ subjective ratings, is more influenced by experts’ preferences and cannot reflect the changes in objective conditions.

The CRITIC method is an objective weight assignment method, which is proposed by Diakoulaki. Its main goal is to establish the indicators’ objective weights using two fundamental ideas. The first is contrast intensity, which reflects, in terms of standard deviation, the size of the difference between the values obtained by each assessment scheme for the same index. The size of standardized deviation indicates the size of the gap between values taken by schemes within the same indicator, and the more standardized deviation, the more significant the difference in values taken by the various schemes. The second is the conflict between the evaluation index based on the correlation. If two indexes have a significant positive correlation, their conflict is low. The CRITIC method also uses the coefficient of variation to make the dispersion reflected by the standard deviation more relevant to reality on the basis of considering the amount of information of the indicators and factor correlation, which has significant superiority [43]文献32. In contrast, other objective weight assessment methods, such as the entropy weighting method and the TOPSIS method, without considering the interaction and interdependence between indicators, can thus not deal with the multi-index information aggregation assessment.

Therefore, this paper uses the procedure of calculating the information quantity of indicators in the improved CRITIC theory to replace the process of G1 method indicator assignment, and uses the information quantity of indicators as the indicator assignment to construct a comprehensive assignment method of the improved CRITIC-G1 method, which not only retains the importance ranking of indicators but also reflects the size of the information quantity contained in indicators, making the assignment results more objective and reasonable [44]. The precise steps of the calculation are listed below.

Step 1. Importance ranking. Five professionals with expertise in the digital village were invited to rank the importance level of evaluation indicator i based on their own experience in the field.

The improved CRITIC method determines the objective weight of indicator i by calculating the product of the contrasting intensity within the evaluation indicator i itself and the conflicting product between each other indicator. If the more information content of indicator i has, the more information it has and transmits, the greater the impact on the evaluation object and the greater the weight.

Step 2. Determine the comparison intensity of the evaluation index i itself c_v(i), dividing the standard deviation of indicator i by its mean value. (1) Where σ_(i) is the standard deviation of the i th assessment index; μ_(i) is the mean value of the i th assessment index.

Step 3. Calculate the conflict between evaluation indicators Cf(i). (2) Where r_ki is the correlation coefficient between assessment index k and i.

Step 4. Calculate the amount of information C_i.

(3)

Step 5. Calculate the ratio of importance degree r_i. Determine the importance ratio of adjacent indicators by calculating the information content of indicators.

(4)

Step 6. Determine the evaluation index weights. According to the ratio of importance r_i,r_i−1,….r₁, the weight of the i th indicator to the assessment object is obtained. (5) Where i is the total number of tertiary indicators under a secondary indicator.

From the weight w_i, the weights of the indicators i-1,i-2,…,1 can be obtained.

(6)

4.2 Extension matter element model

Topology is the law and method to study the possibility of expanding things and to explore and innovate, which is applied to solve contradictory problems. The fundamental goal of matter element analysis is to use ordered triples to represent the three components of things: things, characteristics as well as magnitude, and this triplet is called matter element. Topology element evaluation focuses on evaluating the feasibility and optimality of the solution. Based on the topology theory, the evaluation scheme is taken as the object element, and the evaluation criteria are determined through the qualitative analysis of the topology of the object element, and then the correlation function is introduced for the quantitative calculation to get the affiliation degree of each evaluation scheme to the evaluation criteria, which combines the qualitative and quantitative analysis and takes into account the compatibility of the evaluation indexes. The consequences can be measured more scientifically and rationally [45]. The main steps are as follows.

Step 1. Determine the classical domain element matrix. (7) Where the digital village development is divided into j assessment levels, denoted by Q_j, the classification of levels depends on the specific situation; the set of indicators is denoted by V (v₁, v₂,…,v_n), including a total of n indicators; the classical domain C_ij is the range of data of the assessed actual values of the indicator set v_i under different assessment levels Q_j.

Step 2. Determine the nodal domain element matrix. (8) Where the nodal domain C_ip is the range of data of the assessed real values of the assessment indicator set V under the whole assessment level Q.

Step 3. Determine the elements to be evaluated. (9) Where C (c₁, c₂,…,c_n) denotes the evaluation data values of the elements to be evaluated under different indicators.

Step 4. Calculate the posting progress of the element to be evaluated with each grade K_j(R₀). (10) (11) Where D_j(v_i) and w_i(X) refer to the separation of the element to be assessed from the classical domain and the size of the indicator weight, respectively.

Step 5. Determine the class to which it belongs. If there is material element posting progress to be evaluated K_j’(Q₀) = max{K_j(Q₀)} (j = 1,2,3,4), then it can be determined that the element to be evaluated R₀ belongs to class j’.

5.1 Case description

Jiangxi Province is made up of eleven prefecture-level cities, situated in the middle and lower portions of the Yangtze River in southeast China. It has a prime position and easy access to transportation. Jiangxi Province released "Opinions on the Implementation of Digital Countryside Development Strategy in Jiangxi Province " in 2020 to fully implement the digital countryside development strategy, increase the application of informatization in rural regeneration to its fullest potential and drive the modernization of agriculture and rural areas as a whole, carrying out the spirit of the "Digital Countryside Development Strategy Outline". The "Three-Year Action Plan for the Construction of Digital Agriculture and Rural Areas in Jiangxi Province" was released by Jiangxi Province in 2021. It outlined plans for the establishment of the digital countryside from 2021 to 2023 and created a number of objectives and actions to raise the level of digital growth in Jiangxi Province’s rural districts. It is thus evident that Jiangxi Province maintains a high priority on the digital rural strategy. Jiangxi Province’s digital countryside development level is 42.0%, depending on the "China Digital Countryside Development Report" published in 2022, which surpasses the national average and is located in eighth place, serving as a proxy for China’s overall state of digital rural development. Therefore, this article takes Jiangxi Province of China as an example and collects relevant data for the last six years, from 2016 to 2021, to make a comprehensive evaluation of digital countryside advance in Jiangxi Province from the provincial level. The China Statistical Yearbook and Jiangxi Provincial Statistical Yearbook are the primary sources of research data in this article. The interpolation method is used to fill in for the issue of some specific indicators having missing content in some years.

5.2 CRITIC-G1 empowerment and evaluation of topological elements

Step 1. Determine the index weights. Depending on the digital countryside development index system (see Table 1), the index weights are determined by the method of improving CRITIC correction G1, and the index system weights of this paper are shown in Table 2, as shown in Fig 4.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Indicator weighting chart.

https://doi.org/10.1371/journal.pone.0303847.g004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Calculation results of indicator weights.

https://doi.org/10.1371/journal.pone.0303847.t002

It can be seen that the weight of D2 rural power generation equipment capacity, D4 rural road mileage, D6 number of village committees, D9 minimum number of rural residents living security, D12 digital economy index and D14 rural residents mobile phone ownership is significant, indicating that they are the main factors affecting the evaluation of digital rural development. At present, the rural infrastructure construction in Jiangxi Province is relatively weak, and the living conditions of villagers are not concerned enough, so these indicators are the key factors to consider when evaluating. The digital economy can enable rural industries to achieve digital transformation, improve production efficiency and quality, and increase farmers’ income, which plays a vital role in developing the digital countryside. Secondly, D8 the number of participants in rural social endowment insurance and D10 the total length of postal delivery routes in rural areas have less weight. In recent years, Jiangxi Province attaches great importance to the issue of residents’ pension insurance and implements the national insurance program, so the issue of rural residents’ pension insurance participation has achieved positive results, resulting in a substantial increase in the number of participants. Due to the rapid rise of e-commerce, the limitations of traditional business models have been broken so that rural areas can enjoy the same convenient shopping life as urban areas. Therefore, the impact of the two indicators on the development of the digital countryside is relatively tiny.

Step 2. Classify the evaluation levels. According to the relevant literature, after whole discussion by experts, it was determined that the evaluation level of digital countryside advance was divided into four levels, with level Ⅰ indicating "unqualified", level Ⅱ indicating "qualified", level Ⅲ indicating "good " and level IV indicating "excellent". The corresponding threshold values are displayed in Table 3.’

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Development-level assessment criteria.

https://doi.org/10.1371/journal.pone.0303847.t003

Step 3. On the basis of the criteria in Table 3, determine the classical domain matrix R_j, the nodal domain matrix R_Q and the elements to be evaluated R₀ respectively by Eqs (7)–(9).

Step 4. Determine the posting progress. From the Formula (10), the distance of 21 indicators about four levels of Ⅰ, Ⅱ, Ⅲ, and Ⅳ can be calculated, and the results are displayed in Table 4.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Distance of indicators on the four assessment levels.

https://doi.org/10.1371/journal.pone.0303847.t004

Step 5. Determine the assessment levels. The development assessment of each aspect can be calculated from the Formula (11) and the data in Table 4 about the posting progress of different levels as well as the comprehensive posting progress, so as to determine the evaluation level of each aspect as well as the comprehensive evaluation level, as presented in Table 5.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Correlation and evaluation grade of various aspects of digital village development in Jiangxi Province.

https://doi.org/10.1371/journal.pone.0303847.t005

As seen in Table 5, the comprehensive development evaluation of the digital village in Jiangxi Province is at a good level. Since K_IV>K_II, it indicates that the digital village in Jiangxi Province is developing in an excellent direction and has an increasing development trend. From different aspects, the digitalization of infrastructure, services, economy, and green production is well developed, and the digitalization of life is at an excellent level, which is more prominent. At the same time, infrastructure is developing in an excellent direction. However, in the digitization of services, economy, and green production, there is a trend towards a qualified direction, which is a downward trend. Therefore, it is necessary to identify the reasons for the downward trend in the digitization of services, economy and green production and then formulate targeted measures to strengthen their development. Then, the government continues to grasp the current strong momentum of digital development of infrastructure and life to improve the overall development level of the digital countryside in Jiangxi Province.

5.3 Sensitivity analysis

The evaluation method should have good robustness, and in order to confirm the robustness of the assessment procedure, sensitivity analysis is performed. In practical problems, deciding its assessment indicator system and the weights of each index with the help of expert experience is a simple and effective method, but the reliability of the assessment is unavoidably impacted by human subjectivity. As a result, in this article, the standard weights of each indicator are modified by ±10%, ±20%, and ±30%, respectively, to simulate the changes under different weight cases and compare the effects of these changes on the final evaluation results, as shown in Figs 5–9.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. The weight change rate of tertiary index C1.

https://doi.org/10.1371/journal.pone.0303847.g005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. The weight change rate of tertiary index C2.

https://doi.org/10.1371/journal.pone.0303847.g006

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. The weight change rate of tertiary index C3.

https://doi.org/10.1371/journal.pone.0303847.g007

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. The weight change rate of tertiary index C4.

https://doi.org/10.1371/journal.pone.0303847.g008

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 9. The weight change rate of tertiary index C5.

https://doi.org/10.1371/journal.pone.0303847.g009

The weights of the three levels of indicators are modified separately. Finally, it still emerges that the advance of the digital countryside level in Jiangxi Province belongs to rank III, which is at a good development level. Figs 5 to 9 depict how the optimal posting progress (which is rank III) has changed in terms of absolute change rate with the change of the indicator weights. It is clear that each indicator’s absolute average change rate value has a different change rate value and is essentially symmetrically distributed with the weight change rate 0 at its center. This distribution also shows an approximately linear increase with increasing absolute weight change rate values. The absolute average change rate value is roughly the same when the absolute value of the weight change rate for a specific assessment factor is the same, meaning that it is similarly sensitive to the assessment outcomes. Meanwhile, it can be seen that D4 has the most enormous sensitivity to the weights, and the largest absolute mean rate of change value is 0.0329%, which is significantly less than the weight change rate size, demonstrating that assessment outcomes are somewhat steady. Therefore, the evaluation method has good robustness.

5.4 Comparative analysis

As a way to verify the rationality and accuracy of the proposed method, the AHP-entropy weighting approach and the coupling degree assessment model were applied to assess the advance of digital countryside in Jiangxi Province, and the results were good, which demonstrated the validity of the suggested approach.

In comparison to the AHP-entropy weighting method, the improved CRITIC modified G1 weighting method used in this paper retains the importance ranking of indicators and reflects the information size contained in the indicators, taking into account the mutual influence of indicators, and does not require consistency testing, which is simpler and more accurate in determining the weights of indicators. Compared with the coupling degree evaluation method, the topological element method used in this paper focuses more on the feasibility and optimization of the evaluation scheme, introducing the correlation degree function for quantification, and the calculation entirely takes into account the compatibility of the evaluation indicators, which is more operable, realistic and has more practical value.