Impact of Digital Literacy on Farmers’ Adoption Behaviors of Green Production Technologies

Abstract

The application of digital technology offers new opportunities to promote the green transformation and upgrading of agriculture. Farmers’ digital literacy, as a critical link between digital technology and agricultural green development, significantly influences their production decisions. Whether digital literacy serves as an enabling factor driving farmers’ adoption of agricultural green production technologies warrants further exploration. This paper uses the entropy method to measure farmers’ digital literacy levels and employs a Probit model for empirical analysis of survey data from 643 farmers in Shandong and Shaanxi provinces, examining how farmers’ digital literacy influences their adoption of green production technologies. The baseline regression result indicates that digital literacy can significantly increase farmers’ adoption of green production technologies. A mechanism analysis reveals that enhanced farmers’ digital literacy promotes the adoption of green production technologies through three pathways: enhancing farmers’ risk perception, expanding farmers’ digital social capital, and strengthening the effectiveness of technology promotion. Heterogeneity analysis demonstrates that improved digital literacy significantly enhances the adoption of four technologies—water-saving irrigation, pest control, pollution-free pesticide, and straw return to fields—and exerts a stronger impact on large-scale and middle-generation farmers. Accordingly, this study suggests improving digital village infrastructure, enhancing farmers’ digital literacy comprehensively, and formulating differentiated extension policies.

1. Introduction

In recent years, China’s grain production has increased steadily, reaching historic milestones. According to the 2024 China Statistical Yearbook, by the end of 2023, China’s total grain production reached 695.41 million tonnes. However, the resource-intensive agricultural production model, which places excessive emphasis on maximizing yields, has resulted in the overuse and inefficient application of chemical fertilizers and pesticides. This practice has not only led to widespread soil contamination but also exacerbated agricultural non-point source pollution, which refers to the dispersion of pollutants from agricultural activities into the farming environment, including fertilizer and pesticide pollution. These issues pose significant barriers to sustainable development and have long-term negative impacts on the environment and ecosystems. In response, the Chinese government introduced the Technical Guidelines for Green Agricultural Development (2018-2030), emphasizing that establishing a green agricultural development technology system is key to implementing sustainable development strategies and addressing pressing resource and environmental issues in agriculture and rural areas. However, traditional agricultural information dissemination channels, along with farmers’ resource constraints and cognitive limitations, have hindered their adoption of green production technologies, leaving farmers with insufficient intrinsic motivation. In this context, promoting the shift from conventional to green production among farmers, particularly establishing a sustainable mechanism for the adopting green production technologies, has become an urgent issue.

Research by domestic and international scholars on farmers’ behavior and willingness to adopt green production technologies mainly focuses on three key aspects. First, the existing literature, based on farmers’ resource endowment characteristics, primarily examines how economic, human, and social capital influence their behavioral decisions. Economic capital includes factors like annual income, cultivated land area, and financing ability [1,2]; human capital encompasses labor force population, educational level, and health status [3]; and social capital involves intangible resources such as social network, Information transmission, and interpersonal trust [4]. Second, from the perspective of farmers’ endogenous drivers, existing research highlights factors such as cognitive ability, risk perception, and information acquisition [5,6]. Third, external environmental factors, including agricultural social services, cooperative participation, technology promotion, and institutional constraints, significantly affect farmers’ behavioral decisions [7,8,9,10].

The widespread adoption of digital technology and the rapid growth of the digital economy are driving a significant transformation in China’s rural areas. Advances in agricultural information technologies—such as big data, cloud computing, Mobile Internet, the Internet of Things, and artificial intelligence—have driven the growth of the agricultural digital economy [11]. Data from the Ministry of Industry and Information Technology (MIIT) show that the number of rural broadband users reached 195.31 million in April 2024. Additionally, by December 2023, rural Internet users totaled 326 million, as stated in the 53rd Statistical Report on the Development of the Internet in China. These figures highlight significant improvements in rural information infrastructure, providing a strong foundation for the extensive application of digital technologies in agriculture. In the digital economy era, digital literacy is a critical measure of farmers’ ability to adapt to modern agricultural development. Digital literacy generally refers to an individual’s ability to define, access, manage, integrate, disseminate, evaluate, and create information securely and effectively using digital technologies and Internet-enabled devices in economic and social contexts. For farmers, digital literacy not only encompasses basic information acquisition and processing skills but also emphasizes their application in agricultural production and management. The measurement of farmers’ digital literacy typically includes five dimensions: information and data literacy, communication and collaboration literacy, digital content creation literacy, digital security literacy, and problem-solving literacy [12,13,14]. Through the assessment of these dimensions, farmers’ ability to obtain and apply agricultural information, make technological decisions, and participate in online agricultural training can be comprehensively evaluated [15]. In recent years, China has made significant progress in improving the digital literacy of its population. According to the 2024 National Digital Literacy and Skills Development Survey Report, over 60% of Chinese citizens possess digital literacy and skills at a basic level or higher, with digital literacy levels correlating with regional economic development. In rural areas, 50.57% of rural adults have digital literacy at the basic level or above, and 9.53% have advanced digital literacy. Among rural minors, 53.11% possess basic or higher digital literacy, with 6.33% at an advanced level. Additionally, 79.06% of agricultural production and auxiliary personnel have basic or higher digital literacy, with 33.77% at an advanced level. Improving digital literacy helps farmers more effectively access and analyze global market information, optimize production and sales decisions, and promote the adoption of new technologies, thereby enhancing productivity and market competitiveness. Furthermore, digital literacy increases the transparency of market information, reduces information asymmetry, and helps farmers mitigate risks and improve their income [16]. Only individuals or groups with high digital literacy can effectively utilize digital technologies [17], enabling them to seize opportunities in the wave of agricultural green transformation and achieve sustainable agricultural development.

While some scholars have acknowledged the influence of digital literacy on farmers’ adoption of green production technologies [18], few studies have explored the underlying mechanisms in depth. Theoretically, digital literacy can help farmers overcome barriers to using digital technology, improve their ability to access information, acquire skills, and emulate practices, thereby supporting decisions like adopting agrotechnologies [19]. Second, adopting green production technologies requires farmers to have adequate information [20]. Enhanced digital literacy reduces information asymmetry, enabling farmers to fully understand and adopt these technologies [21]. Additionally, improving digital literacy contributes to human capital accumulation. As farmers’ human capital increases, their production behaviors and concepts change significantly, facilitating the adoption of green production technologies [22]. Finally, digital literacy enhances farmers’ understanding of green production technologies and supports their adoption through digital agricultural extension platforms, including agricultural apps and public accounts [23]. Furthermore, existing studies often measure digital literacy in a narrow manner, primarily referencing UNESCO’s Global Digital Literacy Framework [24], and lack comprehensive assessments of core digital competencies at the individual farmer level.

Based on this, the purpose of this study is to analyze the influence of digital literacy on farmers’ adoption of green production technologies using micro-survey data collected from farmers in Shaanxi and Shandong provinces in 2022. Specifically, this study aims to achieve the following three research objectives. First, the paper establishes a digital literacy evaluation index system based on five dimensions: information and data literacy, communication and collaboration literacy, digital content creation literacy, digital security literacy, and problem-solving literacy. The entropy method is then used to measure the digital literacy levels of the surveyed farmers, providing a micro-level analysis of the digital literacy status of rural residents. Second, the paper employs a Probit model to analyze the overall impact of digital literacy on farmers’ adoption of green production technologies and uses a mediation effect model to examine the mechanisms through which digital literacy influences the adoption of these technologies, specifically through risk perceptions, digital social capital, and technology promotion. Third, this study uses subgroup regression to differentiate the effects of digital literacy on various types of green production technologies, as well as its impact on farmers with different farm sizes and generations. This approach aims to provide a more comprehensive analysis of the tendencies in decision-makers’ technology adoption behaviors. Additionally, due to China’s national context of a large country with small-scale farmers, our study focuses on small farmers as the research subjects. By incorporating digital literacy into the analytical framework, this study deepens the understanding of the mechanisms behind farmers’ adoption of green production technologies in the digital age. It expands the scope and connotation of digital literacy, particularly focusing on its role in agricultural green development. By refining digital literacy into multiple dimensions and exploring its driving effects on farmers’ adoption of green technologies, this paper enhances the theoretical framework of digital empowerment for agricultural green development, offering new perspectives and enriching the theoretical research on agricultural green development.

2. Theoretical Analysis and Research Hypotheses

The diffusion of innovation theory suggests that the technology adoption process is influenced by several factors, including communication channels, relative advantage, compatibility, complexity, trialability, and observability [25]. First, digital literacy enhances farmers’ risk perception. According to the concept of trialability in the diffusion of innovation theory, when farmers can experiment with new technologies on a small scale, their perceived risk is reduced. Digital literacy enables farmers to access more information about green production technologies through digital platforms and tools, thereby reducing uncertainty and increasing their willingness to adopt the technology and their capacity to manage risks. Second, digital literacy facilitates communication and interaction among farmers by fostering digital social capital, thereby enhancing the observability and compatibility of the technology. The theory emphasizes that the greater the observability of an innovation, the more rapid its diffusion. Farmers share their experiences, feedback, and results regarding green production technologies through social media, online forums, and digital platforms, thereby increasing other farmers’ recognition and trust in the technology, which accelerates its spread and adoption. Lastly, digital literacy enhances the effectiveness of technology promotion. According to the diffusion of innovation theory, the diffusion of an innovation is influenced by effective communication channels. Digital literacy enables farmers to more effectively access information from the government, agricultural experts, and extension agencies, and engage in online training and remote education via digital platforms, thereby promoting the widespread application and rapid diffusion of green production technologies. Based on the above analysis, the paper proposes the first hypothesis.

Hypothesis 1 (H1).

Enhancing digital literacy facilitates farmers’ adoption of green production technologies.

Prospect theory suggests that individual decision-making is influenced by a combination of risk perception and subjective judgment. Farmers’ risk aversion may hinder the adoption of new agricultural technologies [26]. While green production technologies are crucial for reducing chemical pesticide use and promoting sustainable agricultural development, their adoption by farmers also entails risks, including uncertain net returns and the potential for misuse. Farmers with higher digital literacy are better equipped to assess the benefits and risks of new technologies and are more willing to experiment with green production techniques. On the one hand, Internet use breaks down information access barriers, enhancing the information available to farmers [27]. Enhanced digital literacy allows farmers to efficiently access abundant information on green production technologies via the Internet, agricultural apps, online forums, and other sources. The availability of information reduces farmers’ fear and uncertainty about new technologies, thereby increasing their risk tolerance for adopting unknown technologies. On the other hand, farmers’ risk attitudes directly influence their willingness to adopt new technologies and their adoption behaviors [28,29]. When farmers’ risk appetite increases, they are more likely to explore and experiment with new green production technologies, rather than adhering to traditional methods. This willingness to explore is a key driver of technology adoption. The higher farmers’ risk perception, the more they focus on the long-term environmental and economic benefits of green technology in its early adoption phase, reinforcing their commitment to sustainable development and increasing their enthusiasm for adopting green production technologies. Based on the above analysis, the paper suggests that the impact of digital literacy on farmers’ adoption of green production technologies can be mediated by the intermediary variable of risk perception. Therefore, the paper proposes the second hypothesis.

Hypothesis 2 (H2).

Enhancing digital literacy increases farmers’ risk perception, thereby promoting the adoption of green production technologies.

The theory of social capital emphasizes the role of social relationships, networks, and norms in facilitating both individual and collective action. Digital social platforms offer farmers abundant information resources and social networks, boosting their ability to access external support and reducing uncertainty in adopting green production technologies. Enhancing farmers’ digital literacy can increase their use of social media, such as WeChat, overcome the geographic limitations of social networks [30], improve communication frequency between farmers and technology suppliers, and strengthen existing social network relationships. On the one hand, the social attributes of the Internet can reshape farmers’ interpersonal networks and foster the accumulation of social capital [31]. The Internet helps farmers overcome geographical and identity limitations, significantly reducing the cost of establishing social relationships. Digitally literate farmers can use various online platforms to establish new types of weak ties, thereby expanding their social network scale [32]. On the other hand, social networks offer farmers diverse information access channels and can effectively transmit information on agricultural technology [33]. Social networks help farmers obtain green production technology information from various channels and levels, promote interactive communication, and encourage support among network members. This enhances the speed of information dissemination, alleviates information asymmetry, reduces transaction costs in green production technology adoption, and positively impacts farmers’ adoption behavior [34]. Based on the above analysis, the paper suggests that the impact of digital literacy on farmers’ adoption of green production technologies can be mediated by the intermediary variable of digital social capital. Therefore, the paper proposes the third hypothesis.

Hypothesis 3 (H3).

Enhancing digital literacy expands farmers’ digital social capital, thereby promoting the adoption of green production technologies.

Unlike the homogeneous information from direct neighborhood exchanges, the heterogeneous information provided by technology promotion through digital platforms also facilitates farmers’ adoption of green production technologies. The higher farmers’ digital literacy, the better their access to and understanding of information, enabling them to more effectively receive government-promoted information. Additionally, the use of digital tools (e.g., mobile apps and online seminars) enhances the efficiency and coverage of government efforts to promote green production technologies. On the one hand, farmers need adequate learning ability and digital literacy to obtain valuable information via digital agricultural extension apps [35]. As 5G and other digital infrastructures continue to develop in rural areas, the Internet has become a platform for sharing agricultural technology information, overcoming time and space constraints and enabling farmers to access agricultural information and technical guidance from home [36]. On the other hand, technology promotion is generally considered a key determinant when exploring the factors influencing farmers’ technology adoption [37]. Studies have analyzed the impact of technology promotion on farmers’ technology adoption in terms of both breadth and depth. Specifically, the reliability of government technology extension helps reduce farmers’ resistance and adoption risks, while extension workers enhance farmers’ knowledge and provide agricultural technology services [38]. Agricultural technology training improves farmers’ cognitive abilities, learning capacity, and practical skills, reduces barriers to technology use, alleviates information asymmetry, and allows farmers to use Internet tools to communicate easily with agricultural technicians and receive more guidance on agricultural technology [39]. With the development of digital technology, technology promotion has evolved from traditional methods to a combination of conventional and digital agricultural technology extension. Digital agricultural technology services broaden farmers’ information access channels, improve efficiency, and deepen their understanding of green technology [40]. By visualizing and facilitating the two-way flow of information, these services help farmers overcome barriers to mastering technical information and cover all aspects of agricultural production. They enable farmers to receive timely and effective technical guidance in daily operations, directly influencing their adoption behavior through precise matching of production links [41]. Based on the above analysis, he paper suggests that the impact of digital literacy on farmers’ adoption of green production technologies can be mediated by the intermediary variable of technology promotion. Therefore, the paper proposes the fourth hypothesis.

Hypothesis 4 (H4).

Digital literacy enhances the effectiveness of technology promotion, which, in turn, facilitates the adoption of green production technologies by farmers.

Based on the above assumptions, this paper constructs a theoretical analysis framework diagram, as shown in Figure 1. That is, digital literacy can influence farmers’ adoption behavior of green production technologies through risk perception, digital social capital, and technology promotion.

3. Materials and Methods

3.1. Data and Sample Characteristics

The data used in this study were collected from a household survey conducted by the research team from July to August 2022 in Shaanxi and Shandong provinces, with the research area shown in Figure 2. The main reasons for selecting these two provinces are as follows: First, the adoption of green production technologies is representative. By the end of 2022, Shandong and Shaanxi provinces had a grain sown area of 8.37 million hectares and 3.01 million hectares, respectively, and grain output of 55.43 million tonnes and 12.97 million tonnes, respectively, both of which had a higher demand for green production technologies. Both Shaanxi and Shandong have national pilot zones for the development of green production technologies in agriculture, making them well-suited for studying the adoption of these technologies by farmers. Second, the level of digital development in agriculture and rural areas is typical. The two provinces, representing the eastern and western regions of China, have achieved remarkable progress in promoting the digital transformation of agriculture and rural areas. Both provinces have committed to expanding broadband networks, improving mobile communication facilities, and constructing data centers in rural areas, laying a solid hardware foundation for applying digital technology in agriculture. According to data from the National Bureau of Statistics and the statistical yearbooks of Shandong and Shaanxi, by the end of 2023, the number of Internet broadband access ports in Shandong and Shaanxi provinces was 74.38 million and 30.08 million, respectively, and the number of rural broadband access subscribers was 11.48 million and 5.34 million, respectively. By the end of 2022, the average number of mobile phones per 100 rural households in Shandong and Shaanxi provinces was 235.2 and 265.6, respectively, with higher Internet penetration rate and better digital infrastructure, making them typical for studying the digital literacy of rural households.

The specific process involved randomly selecting 1–2 cities from each province, 1–2 districts and counties from each city, 2–4 townships from each district and county, 4–5 villages from each township, and finally 25–30 farmers from each village as research subjects. In the field survey, the research team flexibly adjusted the number of participants based on the specific conditions of each survey. In order to ensure the authenticity of the data and consistency in operations, the research team refined the questionnaire based on a preliminary survey and provided multiple rounds of training for the interviewers. During the actual survey, one-on-one home visits were conducted. The research distributed 684 questionnaires. After collating the responses and excluding those with missing data, inconsistencies, or outliers, 643 valid questionnaires were obtained, resulting in a validity rate of 94%. The samples were well-represented. The questionnaire survey was conducted through one-on-one interviews, focusing mainly on individual and household characteristics of farmers, agricultural production and management status, digital literacy levels, and the adoption behaviors of green production technologies.

The basic characteristics of individual farmers, households, and policies in the survey sample are presented in

Table 1. Results of the descriptive statistical analysis of the sample.

Categorical Variable		Sample Size	Proportion (%)	Categorical Variable		Sample Size	Proportion (%)
annual income (10,000 yuan)	[0, 5)	195	30.33%	region	Xianyang	183	28.47%
	[5, 10)	216	33.59%		Baoji	295	45.86%
	[10, ∞)	232	36.08%		Liaocheng	165	25.67%
land size (mu)	[0, 5)	390	60.65%	gender	male	589	91.59%
	[5, 10)	103	60.65%
	[10, ∞)	150	23.33%		female	54	8.41%
age	[0, 45)	67	10.42%	political identity	yes	75	11.65%
	[45, 60)	260	40.44%
	[60, ∞)	316	49.14%		no	568	88.35%
years of education	[0, 5)	85	13.22%	government regulation	yes	43	6.69%
	[5, 9)	258	40.12%
	[9, ∞)	300	46.66%		no	600	93.31%
labor force population	0	65	10.11%	technology promotion	yes	243	37.84%
	1~2	326	50.70%
	3~4	237	36.86%
	5	15	2.33%		no	400	62.16%

. In the sample regions, 63.92% of farm households had an annual income of less than 100,000 yuan. Additionally, 76.67% of farmers cultivated less than 10 mu of land. Most farm households had 1–2 labor force members, and 40.12% of farmers had 5–9 years of education. Furthermore, 37.84% of farmers had received technical training promoted by the government, 11.65% of farmers were party members, and 49.14% of farmers were aged 60 or above. The specific descriptive statistics results can be found in Table 1.

3.2. Research Methodology

First, this study calculates the digital literacy levels of farmers in the sample area based on survey data. Through prior analysis, farmers’ digital literacy is categorized into five dimensions. Given the need for an objective and scientific weighting method, this study employs the entropy weight method to assign weights to these five dimensions, thereby calculating the overall digital literacy level of farmers. Second, since the dependent variable, the adoption behaviors of green production technologies, is a binary variable, this study uses the Probit model to examine the direct impact of digital literacy on farmers’ adoption behaviors of green production technologies. Finally, to explore the mechanism through which digital literacy influences farmers’ adoption behaviors of green production technologies, this study employs a mediation model to analyze the indirect pathways through which digital literacy affects farmers’ technology adoption behaviors.

3.2.1. Entropy Method

In this paper, the entropy method is used to assign weights to the various indicators of digital literacy, and the comprehensive digital literacy level of farmers is then derived using the weighted average method [42]. The entropy value method is an objective approach used to determine the weight of each evaluation index. The specific steps are as follows:

(1) Data standardization: The raw data were standardized to eliminate the effects of scale differences between indicators. $X_{ij}$ represents the original value of the ith farmer on the jth indicator, and $max\left(x_j\right)$ and $min\left(x_j\right)$ denote the maximum and minimum values of the jth indicator, respectively. The formula is as follows:

$$X_{ij}^{*}={\displaystyle \frac{x_{ij}-min\left(x_j\right)}{max\left(x_j\right)-min\left(x_j\right)}}$$

(1)

(2) Calculation of indicator entropy: The entropy value was calculated for the normalized data. Here, $e_j$ represents the entropy value of the jth indicator, $n$ denotes the number of farmers in the sample, and $p_{ij}$ represents the weight of the ith farmer on the jth indicator. The formula is as follows:

$$e_j=-{\displaystyle \frac{1}{ln\left(n\right)}}{\sum }_{i=1}^n{p}_{ij}ln\left(p_{ij}\right)$$

(2)

$$p_{ij}={\displaystyle \frac{x_{ij}^{*}}{{\sum }_{i-1}^n{x}_{ij}^{*}}}$$

(3)

(3) Calculation of the coefficient of variation, $g_j$, for each indicator. The formula is as follows:

$$g_j=1-e_j$$

(4)

(4) Determination of the weights of the indicators. Based on the coefficient of variation, the weight of each indicator, denoted as $w_j$, is calculated, where a higher weight indicates that the indicator provides more information, and m represents the total number of indicators. The formula is as follows:

$$w_j={\displaystyle \frac{g_j}{{\sum }_{j-1}^m{g}_j}}$$

(5)

(5) Comprehensive measurement of digital literacy level. The digital literacy indicators are weighted and summed using the determined weights, where $S_i$ represents the level of digital literacy of the ith farmer. The formula is as follows:

$$S_i={\sum }_{j=1}^m{w}_j{x}_{ij}^{*}$$

(6)

3.2.2. Probit Model

In this paper, a Probit model is constructed to test the effect of digital literacy on farmers’ adoption of green production technologies [43]. The formula is as follows:

$${adopt}_i={\beta }_0+{\beta }_{1}D{L}_i+{\beta }_2{X}_i+{\epsilon }_i$$

(7)

In this model, ${adopt}_i$ represents the adoption behaviors of green production technologies, $D{L}_i$ represents the level of digital literacy of farmers, $X_i$ denotes control variables, and ${\epsilon }_i$ is the error term. Considering the potential issue of heteroskedasticity in cross-sectional data, which may lead to biased parameter estimates, robust standard errors are employed in all regression analyses. The parameter ${\beta }_1$ in this model indicates the effect of farmers’ digital literacy level on the adoption of green production technologies.

3.2.3. Intermediation Model

This paper uses stepwise regression [44] to examine the mediating roles of risk perception, Internet-based social capital, and technology promotion in the relationship between digital literacy and the adoption of green production technologies by farmers, as outlined in the following steps:

$$Y=cX+e_1$$

(8)

$$M=aX+e_2$$

(9)

$$Y=c^{\prime }X+bM+e_3$$

(10)

In Equations (8)~(10), $M$ represents risk perception, digital social capital, and technology promotion; $c$ represents the total effect of digital literacy on farmers’ adoption of green production technologies; $a$ represents the effect of digital literacy on risk perception, digital social capital, and technology promotion; $b$ represents the effect of risk perception, digital social capital, and technology promotion on the adoption behaviors of green production technologies by farmers, after controlling for digital literacy; $c^{\prime }$ represents the direct effect of digital literacy on the adoption behaviors of green production technologies, after controlling for the effects of risk perception, digital social capital, and technology promotion; and $e_1$, $e_2$, and $e_3$ are random error terms.

3.3. Variable Selection

3.3.1. Dependent Variable

This paper uses the adoption behaviors of green production technologies as the dependent variable. Based on the “Technical Guidelines for Agricultural Green Development (2018–2030)” and relevant studies [45,46,47], five agricultural production technologies required by farmers in the production process are selected: “water-saving irrigation technology”, “soil-measuring formula fertilizer technology”, “pest control technology”, “pollution-free pesticide technology”, and “straw return technology”. These five technologies serve as examples of agricultural green production technologies. In the survey, the farmers’ adoption of these green production technologies is characterized by whether they have adopted any of the following technologies during planting: water-saving irrigation technology, pollution-free pesticide technology, soil-measuring and formulated fertilizer technology, straw-return technology, and integrated pest control technology. If farmers adopt any of these technologies, they are assigned a value of 1; if none of these technologies are adopted, the value is 0.

3.3.2. Core Independent Variables

This paper takes digital literacy as the core independent variable. In the context of green agricultural production, digital literacy is not only reflected in the operational skills related to digital technologies, but also in critical understanding of information, the absorption and recreation of green production technology knowledge, and the ability to promote technology adoption. In the agricultural sector, digital literacy is one of the key factors influencing farmers’ adoption of green production technologies. A high level of digital literacy enables farmers to efficiently access information on green production technologies, and to communicate and collaborate with technology promoters and other farmers. Based on the theoretical framework of the EU DigComp 2.1 [12,13,14], and taking into account the practical production and business conditions of farmers in the surveyed region, this study constructs a digital literacy evaluation index system with five dimensions: information and data literacy, communication and collaboration literacy, digital content creation literacy, digital security literacy, and problem-solving literacy. The entropy method will be used to calculate the weights of these five secondary indicators after standardizing the relevant indicators, objectively determining the weights of the farmers’ digital literacy indicators. The specific measurement items are shown in

Table 2. Digital literacy measurement framework.

Dimension	Specific Measurement Items	Mean	Std.Dev
information and data literacy	ability to effectively use online tools to browse, search, and filter data and information	3.782	0.889
	ability to assess the credibility of sources of data and information obtained from the Internet	3.491	0.817
	ability to compare information from different websites to ensure the reliability of the obtained information	3.368	0.866
communication and collaboration literacy	proficient in using messaging apps such as WeChat or QQ for online communication with others.	4.063	0.839
	proficient in using digital tools to participate in the management of village public affairs.	3.309	0.92
	utilize appropriate digital technologies to share data and information with others.	3.628	0.824
digital content creation literacy	proficiently use social media to post personal updates or status on social networks.	3.76	0.928
	use digital tools to create and post relevant online short videos or clips.	3.66	0.966
	effectively use digital tools for activities such as learning, entertainment, shopping, or consumption.	3.894	0.952
digital security literacy	effectively use software to detect and remove viruses from electronic devices.	3.363	0.97
	clearly disable the location services on electronic devices.	3.076	0.894
	effectively use software features to protect personal information security (such as setting security questions, facial recognition, etc.).	3.676	0.955
problem-solving literacy	identify and resolve issues encountered during the operation and use of digital devices.	3.239	0.892
	use digital technology tools to solve the practical problems currently encountered.	3.287	0.812
	clearly understand and recognize the areas in which one needs to improve digital literacy.	3.335	0.869

Note: Data source was accessed from July and August of 2022.

, and a five-point Likert scale is used with the options ‘Strongly Disagree = 1, Disagree = 2, Neutral = 3, Agree = 4, Strongly Agree = 5.’

3.3.3. Intermediary Variables

Based on the previous analysis of the theoretical mechanisms of digital literacy’s impact on the adoption of green production technologies by farmers, this paper selects three key intermediary variables: risk perception, digital social capital, and technology promotion. Specifically, ‘farmers’ willingness to take out loans when they are unable to afford them’ is used to characterize farmers’ risk perception; ‘whether farmers have joined public groups for village affairs’ is used to characterize farmers’ digital social capital; and ‘whether the government has publicized green production technologies’ is used to characterize technology promotion.

3.3.4. Control Variables

To control for the influence of other factors on the empirical results, this paper draws on the existing literature [48,49,50] and selects individual characteristics of farmers, household characteristics of farmers, and policy characteristics as control variables. Specifically, for individual characteristics of farmers, the gender, age, educational level, health status, and political identity of household head were selected. For household characteristics of farmers, household size, labor force population, cultivated land area, and annual income were selected. For policy characteristics, government regulation was selected. The specific measurement items are shown in

Table 3. Definition and summary statistics of the selected variables.

Variable Type	Variables	Definition	Mean	Std.Dev
dependent variable	adoption behaviors of green production Technologies	whether green production technologies have been adopted in the planting process: Yes = 1, No = 0.	0.547	0.498
core independent variable	digital literacy	the comprehensive digital literacy score is calculated using the entropy method.	0.62	0.123
intermediary variables	risk perception	willingness to take out a loan when farmers are unable to afford it: Yes = 1; No = 0.	0.663	0.473
	digital social capital	whether the farmer has joined a village affairs public group: Yes = 1; No = 0.	0.848	0.36
	technology promotion	whether the government has promoted green production technologies: 1 = Yes; 0 = No.	0.378	0.485
control variables	gender	gender of household head: Male = 1; Female = 0	0.916	0.278
	age	age of household head	58.732	10.992
	educational level	the number of years of formal education received by the household head	7.647	3.371
	health status	health status of the household head: Very unhealthy = 1; Unhealthy = 2; Average = 3; Relatively healthy = 4; Very healthy = 5.	2.106	1.618
	political identity	whether the household head is a Party member: Yes = 1; No = 0.	0.117	0.321
	household size	the number of permanent residents in the household	3.747	1.656
	labor force population	the number of labor force population in the household	2.309	1.219
	cultivated land area	the land area farmed by the household.	7.862	9.975
	annual income	the actual annual income of the farmer’s household	9.929	9.073
	government regulation	whether the government has conducted sampling inspections on agricultural products: Yes = 1; No = 0.	0.067	0.25

Note: Data source was accessed from July and August of 2022.

.

4. Results

4.1. Basic Regression

To address the issue of heteroskedasticity caused by the random disturbance term, all empirical results in this paper are based on robust standard errors. Models 1 and 2 in

Table 4. Regression results of the models of the impact of digital literacy on farmers’ adoption of green production technologies.

Variables	Model 1	Model 2
digital literacy	0.630 *** (0.156)	0.351 ** (0.153)
gender	-	−0.032 (0.067)
age	-	0.008 *** (0.002)
years of education	-	0.029 *** (0.007)
health level	-	−0.001 (0.015)
political identity	-	−0.021 (0.058)
household size	-	0.044 *** (0.014)
labor force population	-	0.036 * (0.020)
cultivated land area	-	0.004 ** (0.002)
annual income	-	0.000 (0.000)
government regulation	-	0.049 (0.082)
constants	−0.884 *** (0.264)	−3.822 *** (0.557)
regional dummy variable	uncontrolled	controlled
pseudo R2	0.018	0.144
number of observations	643	643

Note: ***, ** and * represent significance levels of 1, 5 and 10%, respectively. Robust standard errors are in parentheses.

show the effect of digital literacy on farmers’ adoption of green production technologies. Model 1 shows that the regression coefficient of digital literacy on the adoption of green production technologies is positive and significant at the 1% statistical level. Model 2 further includes control variables and regional dummy variables, which remain significant at the 5% statistical level, indicating that digital literacy motivates farmers to adopt green production technologies, thereby confirming Hypothesis 1. The regression results for the control variables indicate that the effects of farmers’ age, years of education, household size, and land area on the adoption of green production technologies are significant. Some possible reasons are as follows: Older farmers, with more extensive farming experience, are more likely to recognize the positive impact of green production technologies on agriculture, whereas younger farmers, who are more likely to engage in non-agricultural activities, may be less motivated to adopt green technologies. Higher levels of education are generally associated with greater technical understanding and innovation awareness, making farmers with longer years of education more likely to adopt green production technologies. Household size may reflect the availability of labor resources, with larger households potentially having more labor to implement and manage green production technologies. Farmers with larger land areas typically have more resource endowments to invest in and adopt new green technologies, thereby increasing their willingness to adopt such technologies.

4.2. Endogenous Analysis

The results of the previous empirical analysis suggest that digital literacy facilitates the adoption of green production technologies by farmers. However, the conclusion must account for potential endogeneity issues arising from reverse causality and omitted variables. Specifically, endogeneity issues may arise from two sources. One source is reverse causality, where the relationship between digital literacy and the adoption of green production technologies by farmers is bidirectional. While digital literacy enables farmers to access green production technologies through information access, online learning, and social networks, farmers may also increasingly use digital technologies to seek information and access these technologies due to agricultural production needs, thus further enhancing their digital literacy. Secondly, omitted variables—though this paper includes control variables and regional dummy variables to account for the influence of other factors—may still exist, which affect farmers’ digital literacy and adoption of green production technologies.

To address the potential endogeneity issue, this paper adopts an approach similar to existing studies by using instrumental variable (IV-Probit) modeling. Referring to existing studies [51,52], the instrumental variable chosen is a dummy variable representing whether the Internet is the most important information channel, with a value of 1 if yes and 0 otherwise. This paper uses it as an instrumental variable for digital literacy for several reasons: farmers’ digital literacy is primarily mediated by the Internet, which serves as the prerequisite for improving their digital literacy through mobile phones and computers, showing a strong correlation between the two. Secondly, the Internet being the most important information channel is not correlated with the choice of green production technology, aligning with the exogeneity principle required for instrumental variables.

Table 5. Estimation results of the instrumental variables used in the model of the impact of digital literacy on farmers’ adoption of green production technologies.

Variables	IV-Probit
	First-Stage Regression	Second-Stage Regression
	Digital Literacy	Adoption Behaviors of Green Production Technologies
digital literacy	-	7.312 *** (2.228)
whether the Internet is the most important information channel	0.069 *** (0.011)	-
control variables	controlled	controlled
regional dummy variables	controlled	controlled
first-stage F-statistic	10.91	-
number of observations	643	643
Wald chi2	11.06

Note: *** represent significance levels of 1%. Robust standard errors are in parentheses.

presents the empirical results from the instrumental variable (IV-Probit) model. The Wald exogeneity test in the regression results yielded a ${\chi }^2$ value of 11.06, significant at the 1% level, indicating an endogeneity issue with the digital literacy variable that must be addressed using the IV-Probit model. The first-stage regression results show that the instrumental variable has a significant and positive effect on digital literacy at the 1% level, with an F-value greater than 10, confirming that ’whether the Internet is the most important information channel’ is correlated and not a weak instrument. The second-stage regression results show that the effect of digital literacy on the adoption of green production technology remains positive and significant, consistent with the previous conclusion. The second-stage regression results further confirm that the effect of digital literacy on green production technology adoption remains positive and significant, reinforcing the robustness of the paper’s conclusions after addressing the endogeneity issue with the IV-Probit model.

4.3. Robustness Tests

To further test the robustness of the findings, three robustness testing methods were employed in this study: replacing explanatory variables, adjusting weights, and applying a 1% shrinkage to the variables.

First, the explanatory variables were replaced. Specifically, the dichotomous dummy variable for whether farmers adopt green production technologies was substituted with an ordinal variable representing the degree of adoption. This approach not only indicates whether farmers have adopted the technologies but also captures the extent of adoption. The estimation results are presented in

Table 6. Robustness check.

Variables	Adoption Rate of Green Production Technologies						Equal Weight Method	1% Winsorization of the Variables
digital literacy	0	1	2	3	4	5	0.337 ** (0.151)	0.386 ** (0.153)
	−0.470 ***	−0.002	0.001	0.066 **	0.194 ***	0.212 ***
	(0.138)	−0.001	−0.003	−0.021	−0.058	−0.066
control variables	controlled						controlled	controlled
regional dummy variables	controlled						controlled	controlled
pseudo R2	0.055						0.143	0.146
number of observations	643						643	643

Note: *** and ** represent significance levels of 1 and 5%, respectively. Robust standard errors are in parentheses.

, and the regression results indicate that when farmers adopt no green production technologies, the effect of digital literacy is negatively significant at the 1% level. When farmers adopt 1–2 green production technologies, the effect of digital literacy is not significant. However, when farmers adopt 3–5 green production technologies, the effect of digital literacy becomes positively significant.

Second, the method of measuring digital literacy was replaced. In the benchmark regression, the paper uses the entropy method to calculate the digital literacy level of farmers. To ensure the robustness of the digital literacy indicators, and drawing on practices from existing studies, the paper further applies the equal weight method to recalculate the weights of digital literacy indicators and re-conducts the empirical tests. The regression results, presented in Table 6, show that digital literacy remains significant and has a positive coefficient.

Third, a 1% shrinkage of variables was applied. To exclude the effect of extreme values, the research variables were re-estimated after a 1% shrinkage. As shown in Table 6, digital literacy has a significant and positive coefficient on the adoption of green production technologies. The robustness results consistently show that digital literacy has a significant positive effect on the adoption of green production technologies by farmers, confirming that the benchmark regression results are robust.

4.4. Mechanism Analysis

Previous empirical analyses have shown that digital literacy significantly contributes to the adoption of green production technologies by farmers. However, the pathways through which digital literacy influences this adoption remain unclear. This paper explores the mechanisms of digital literacy’s impact on farmers’ adoption of green production technologies through three pathways: enhancing farmers’ risk perception, expanding farmers’ digital social capital, and enhancing the effectiveness of technology promotion. The specific mechanisms are presented in

Table 7. Estimation results of the mediating effect models of digital literacy on farmers’ adoption behavior of green production technologies.

Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7
digital literacy	0.351 **	0.606 ***	0.281 *	0.499 **	0.321 **	0.319 **	0.268 *
	−0.153	−0.141	−0.153	−0.159	−0.153	−0.151	−0.147
risk perception	-	-	0.116 **	-	-	-	-
			−0.039
digital social capital	-	-	-	-	0.065 *	-	-
					−0.036
technology promotion	-	-	-	-	-	-	0.251 ***
							−0.034
control variables	controlled	controlled	controlled	controlled	controlled	controlled	controlled
regional dummy variables	controlled	controlled	controlled	controlled	controlled	controlled	controlled
pseudo R2	0.144	0.115	0.153	0.057	0.147	0.107	0.196
number of observations	643	643	643	643	643	643	643

Note: ***, ** and * represent significance levels of 1, 5 and 10%, respectively. Robust standard errors are in parentheses.

.

Regressions (2) and (3) in Table 7 show the results of the mediation analysis for the pathway ‘digital literacy → risk perception → adoption of green production technologies’. In regression (2), digital literacy significantly and positively influences farmers’ risk perception at the 1% level, with improved digital literacy enhancing access to information, which, in turn, improves farmers’ risk perception. In regression (3), both digital literacy and risk perception positively and significantly affect the likelihood of adopting green production technologies at the 10% level. This suggests that digital literacy enhances the adoption of green production technologies by increasing farmers’ risk perception, thus confirming Hypothesis 2.

Regressions (4) and (5) in Table 7 present the mediation analysis for the pathway ‘digital literacy → digital social capital → green production technology adoption’. In regression (4), digital literacy significantly and positively influences farmers’ digital social capital at the 1% level, with improvements in digital literacy expanding farmers’ digital social capital through digital network platforms. In regression (5), both digital literacy and digital social capital positively and significantly affect the likelihood of adopting green production technologies at the 5% level. This indicates that digital literacy can promote the adoption of green production technologies by expanding farmers’ digital social capital, thus confirming Hypothesis 3.

Regressions (6) and (7) in Table 7 present the mediation analysis for the pathway ‘digital literacy → technology promotion → green production technology adoption’. In regression (6), digital literacy significantly and positively influences technology promotion at the 5% level, with higher digital literacy among farmers improving the efficiency and timeliness of technology promotion. In regression (7), both digital literacy and technology promotion positively and significantly affect the likelihood of adopting green production technologies at the 10% level. This suggests that digital literacy can enhance the adoption of green production technologies by strengthening the impact of technology promotion, thus confirming Hypothesis 4.

Based on the results of the above mechanism tests, this paper concludes that the positive effect of digital literacy on the adoption of green production technologies by farmers is primarily achieved by enhancing farmers’ risk perception, expanding digital social capital, and enhancing the impact of government technology promotion. Hypotheses 2–4 of this paper are strongly supported.

4.5. Heterogeneity Analysis

4.5.1. Heterogeneity of Green Production Technologies

While the above discussion highlights the positive effect of digital literacy on the adoption of green production technologies, it does not differentiate between different types of green production technologies, potentially obscuring the variation in adoption across different technologies by the sample farmers. By conducting regressions for different types of green production technologies, this paper addresses the above-mentioned issues and provides a more detailed understanding of how various aspects of digital literacy influence the adoption of green production technologies by farmers. This paper classifies green production technologies into five categories: ‘water-saving irrigation technology’, ‘soil-formulated fertilizer technology’, ‘pest control technology’, ‘pollution-free pesticide technology’, and ‘straw return technology’. As shown in

Table 8. Regression results of the model for heterogeneity of green production technologies.

Variables	Water-Saving Irrigation Technology	Soil-Formulated Fertilizer Technology	Pest Control Technology	Pollution-Free Pesticide Technology	Straw-Return Technology
digital literacy	0.269 *	0.148	0.635 ***	0.520 ***	0.264 *
	(0.155)	(0.150)	(0.156)	(0.163)	(0.159)
control variables	controlled	controlled	controlled	controlled	controlled
regional dummy variables	controlled	controlled	controlled	controlled	controlled
Pseudo R2	0.078	0.075	0.134	0.084	0.110
number of observations	643	643	643	643	643

Note: *** and * represent significance levels of 1, 5 and 10%, respectively. Robust standard errors are in parentheses.

, digital literacy has a significant positive effect on the adoption of water-saving irrigation technology, pest control technology, pollution-free pesticide technology, and straw-returning technology. The greatest impact was observed on pest and disease control technologies, followed by pollution-free pesticide technologies, water-saving irrigation technologies, and straw-returning technologies. However, the effect of digital literacy on the adoption of soil-formulated fertilizer technology is not significant, possibly due to various challenges in promoting and applying this technology, such as institutional constraints and resource limitations [53].

4.5.2. Heterogeneity of Cultivated Land Area

In this paper, following existing research [54], the sample was divided into two groups based on the median business size: large-scale and small-scale farmers. Farmers with a land area above the median are classified as large-scale farmers based on the relative scale criterion. Compared to farmers with land area below the median, those above the median possess more cultivated land, reflecting the differences in land scale. This classification is concise and effective, helping analyze the differences in the adoption of green production technologies between large-scale and small-scale farmers. The comparison results in

Table 9. Regression results of the model for heterogeneity of cultivated land area.

Variables	Small-Scale	Large-Scale
digital literacy	0.279 (0.208)	0.343 * (0.204)
control variables	controlled	controlled
regional dummy variables	controlled	controlled
pseudo R2	0.282	0.105

Note: * represent significance levels of 10%. Robust standard errors are in parentheses.

show that, compared to small-scale farmers, the effect of digital literacy on the adoption of green production technologies is positive and significant for large-scale farmers, but not for small-scale farmers. This suggests that enhancing digital literacy is more effective in promoting the adoption of green production technologies among large-scale farmers. A possible explanation is that large-scale farmers possess greater human and economic capital compared to small-scale farmers. Large-scale farmers have higher human capital, a greater ability to utilize digital technologies, and lower marginal costs for using these technologies. Large-scale farmers have higher economic capital and are better equipped to bear the initial investment costs of adopting green production technologies [48], whereas small-scale farmers may be hesitant to adopt such technologies due to limited financial resources and the difficulty in perceiving short-term economic benefits.

4.5.3. Intergenerational Heterogeneity

Agricultural production is typically organized on a household basis, with the household head generally playing a decisive role in decisions regarding the adoption of green production technologies. Drawing on the methods of dividing farm household generations used by other scholars [55], this paper classifies farmers based on their birth years: those born in 1980 or later are classified as younger-generation farmers; those born between 1965 and 1979 are middle-generation farmers; and those born before 1965 are older-generation farmers. The regression results in

Table 10. Regression results of the model for heterogeneity of different generations.

Variables	Older Generation	Middle Generation	Younger Generation
digital literacy	0.125 (0.229)	0.871 *** (0.254)	−0.349 (0.437)
control variables	controlled	controlled	controlled
regional dummy variables	controlled	controlled	controlled
pseudo R2	0.175	0.207	0.293

Note: *** represents significance levels of 1%. Robust standard errors are in parentheses.

indicate that digital literacy significantly influences the adoption of green production technologies among middle-generation farmers, but has no significant effect on younger-generation or older-generation farmers. Based on the survey data from this study, the average farmland area for younger-generation farmers is 5.86 mu, with an average annual income of 12,456 RMB, while the average farmland area for older-generation farmers is 8.46 mu, with an average annual income of 95,949 RMB. These figures suggest that younger-generation farmers do not primarily rely on agricultural production for their income, but are more engaged in non-agricultural activities. As a result, their motivation to adopt green production technologies is relatively low. Meanwhile, due to a lack of innovative inclination and generally lower cognitive abilities, older-generation farmers tend to exhibit a more conservative attitude toward technology adoption. Their lower digital literacy further limits their ability to adapt to new technologies, which results in a lack of motivation and confidence to adopt green production technologies [56].

5. Discussion

First, after conducting field surveys in Shaanxi and Shandong provinces, it was found that both regions have implemented rural digital information infrastructure and disseminate the latest agricultural technology updates and real-time agricultural policy information through “Three Micros and One Terminal” (WeChat, Weibo, micro-videos, and mobile applications). Existing studies have primarily analyzed the factors influencing farmers’ green production behavior from perspectives such as farmers’ resource endowment characteristics [57], technology training [58], and government subsidies [59]. However, discussions from the perspective of farmers’ digital literacy remain insufficient.

Second, digital technology has become a key driving force for the green transformation and upgrading of agriculture, while farmers’ digital literacy plays a crucial role in facilitating the adoption of green production technologies [36]. Accordingly, this study conducts an empirical analysis to investigate the impact of digital literacy on farmers’ adoption of green production technologies. The results show that improving farmers’ digital literacy significantly enhances their adoption of green production technologies, which aligns with the findings of previous studies [60,61], further confirming the importance of digital literacy in driving technology promotion and diffusion and promoting agricultural green development. However, there is limited research that delves into the mechanisms through which digital literacy influences farmers’ adoption of green production technologies. This empirical study finds that farmers’ digital literacy can promote the adoption of green production technologies through three pathways: enhancing farmers’ risk perception, expanding farmers’ digital social capital, and enhancing the effectiveness of technology promotion. This provides an important contribution to the existing literature in this field. Additionally, existing studies often measure digital literacy in a relatively narrow manner, lacking a comprehensive evaluation of the core digital competencies at the individual level. This study analyzes the digital literacy status of farmers from a micro perspective and constructs a digital literacy evaluation index system based on five dimensions: information and data literacy, communication and collaboration literacy, digital content creation literacy, digital security literacy, and problem-solving literacy. This approach extends the research boundaries of farmers’ digital literacy. Further research reveals that the improvement of digital literacy significantly promotes the adoption of four key technologies: water-saving irrigation, pest and disease control, non-hazardous pesticides, and straw returning. The impact is stronger for larger-scale farmers and the middle generation of farmers. The possible reason is that improved digital literacy enables farmers to more efficiently access information related to agricultural technologies, allowing them to select the green production technologies best suited to the local agricultural production environment. Larger-scale farmers, with higher human capital and stronger digital technology usage capabilities, incur lower marginal costs in using digital technologies and are better able to bear the initial investment costs associated with adopting green production technologies. As the dominant group in agricultural production and management, middle-generation farmers are more likely to experiment with and adjust technologies as their digital literacy improves, thereby accelerating the adoption of new agricultural technologies.

Third, it is important to note that, due to the limitations of the micro-level farmer survey and the depth of the issues explored, there is room for optimization in this study. The research is confined to Shaanxi and Shandong provinces, making it difficult to cover other provinces (autonomous regions) in China. Therefore, the generalizability of its conclusions still requires improvement. Future research should focus on expanding the survey to include a wider range of regions and conduct comprehensive investigations of different types of farmers to derive more universally applicable conclusions.

6. Conclusions

Based on data from a micro-survey of 643 farmers in Shaanxi and Shandong provinces, this paper systematically explores the effects and mechanisms through which digital literacy influences the adoption of green production technologies by farmers. The main conclusions are as follows. First, the benchmark regression results show that digital literacy has a significant positive effect on the adoption of green production technologies by farmers. To address potential endogeneity issues arising from reverse causation and omitted variables, this finding has been tested using an instrumental variables approach. Moreover, the empirical results remain valid after a series of robustness tests, including replacing the dependent variable, changing the measure of digital literacy, and applying 1% shrinkage to the variables. Second, the path analysis reveals that improvements in digital literacy significantly enhance farmers’ risk perception, expand their digital social capital, and enhance the effectiveness of technology promotion, thereby influencing their adoption of green production technologies. Third, the heterogeneity analysis shows that digital literacy has a significant positive effect on the adoption of pest control technology, pollution-free pesticide technology, water-saving irrigation technology, and straw-return technology, but no significant effect on the adoption of soil-formulated fertilizer application technology. Moreover, digital literacy is more effective in promoting the adoption of green production technologies among large-scale and middle-generation farmers than among small-scale, new-generation, and older-generation farmers. This study incorporates digital literacy into the analytical framework of farmers’ adoption of green production technologies, deepening the understanding of the mechanisms influencing the adoption of green technologies by farmers in the digital era. It expands the connotations and extensions of digital literacy, particularly in the context of agricultural green development. By refining the multiple dimensions of digital literacy and exploring its driving effects on farmers’ adoption of green technologies, this paper not only enriches the theoretical understanding of digital literacy but also further improves the theoretical framework of digital empowerment for agricultural green development, offering innovative perspectives and profound contributions to the theoretical research in the field of agricultural green development.

7. Policy Recommendations

Based on the above conclusions, this paper presents the following policy recommendations.

First, digital infrastructure in rural areas should be improved. Government departments should continue to strengthen digital infrastructure in rural areas, actively promote the development of gigabit fiber-optic networks, expand network coverage, optimize access to digital resources, increase the frequency with which farmers use digital technology for agricultural information, improve the service capacity of digital facilities, and lower the digital threshold for technology adoption.

Second, the level of digital literacy among farmers should be comprehensively im-proved. This can be achieved through active rural digital education and training initiatives, encouraging research institutes, universities, agricultural enterprises, and other organizations to provide guidance and training to improve farmers’ digital literacy, thereby enhancing their access to technical information via digital technology and supporting the adoption of green production practices. Additionally, a digital platform for the exchange of agricultural green production should be established by integrating the ‘digital network’ with the ‘social network’, facilitating the low-cost dissemination of high-quality technical information and injecting digital vitality into agricultural green production.

Third, differentiated promotion policies should be developed. During the promotion process, it is important to improve the precision of digital literacy policy targeting, focus on key groups, stimulate the intrinsic motivation of new-generation farmers, and leverage their role in bridging the digital divide across generations. At the same time, the differences in capital endowments and demand preferences of farmers should be considered.

8. Limitations of the Article

Despite the valuable insights that this study provides for the surveyed regions, several limitations remain. First, the study utilizes cross-sectional data, focusing on farmers’ technology adoption at the time of the survey, which limits the exploration of the dynamic process of technology adoption. In the future, the research team plans to conduct longitudinal surveys to collect panel data, enabling a more accurate measurement of the evolution of farmers’ technology adoption. Second, this study is concentrated in the provinces of Shaanxi and Shandong in China, meaning the findings may not be directly applicable to regions with different socio-economic conditions, social infrastructure, and cultural contexts. Therefore, caution is needed when generalizing the results to other regions. The research team plans to expand the scope of the survey to include multiple provinces in order to more comprehensively analyze the adoption of green production technologies by farmers across China. Finally, this study may have overlooked potential variables such as cultural attitudes, financial constraints, regional policy differences, infrastructure disparities, and market access. To improve the study, the research team will refine the survey questionnaire and incorporate additional control variables that could affect digital literacy and the adoption of green production technologies, aiming for more precise results.