Sage Journals: Discover world-class research

Abstract

E-commerce entrepreneurship is an important measure for enhancing agricultural efficiency, revitalizing rural areas, and increasing farmers’ income. However, many farmers fall into the dilemma of having the willingness to engage in e-commerce entrepreneurship but failing to take action. Based on empowerment theory, this study constructs an “empowerment–process–outcome” framework and employs survey data from 1,350 farmers in China to empirically analyze the relationship between digital literacy and the intention-behavior gap in farmers’ e-commerce entrepreneurship, along with its underlying mechanisms. The results show that digital literacy is significantly associated with a decreased likelihood of the intention-behavior gap in e-commerce entrepreneurship. Specifically, basic digital literacy is significantly related to behavioral transformation, whereas the practical effects of advanced digital literacy have not yet been fully realized. Furthermore, online social capital can partially mediate this relationship, and both instrumental support and institutional support play important moderating roles. Moreover, heterogeneity analysis reveals that farmers with a preference for cooperative entrepreneurship and those in areas with well-developed rural logistics are more influenced by digital literacy. In addition to enriching the application scenarios of empowerment theory, this study provides insights for optimizing digital literacy development mechanisms and improving the success rate of e-commerce entrepreneurial transformation.

Plain Language Summary

E-commerce entrepreneurship is crucial for boosting agricultural productivity, revitalizing rural regions, and raising farmers’ incomes. However, many farmers express a desire to start e-commerce businesses but fail to put their intentions into action. This study constructs an “empowerment-process-outcome” framework based on empowerment theory. It utilizes survey data from 1,350 Chinese farmers to investigate how digital literacy influences the gap between farmers’ intention to start e-commerce businesses and their actual actions, as well as the underlying reasons for this gap. The findings indicate that an enhancement in digital literacy is associated with a reduction in the gap between intention and action in e-commerce entrepreneurship. Basic digital literacy is strongly tied to the conversion of intentions into actions, while the full benefits of advanced digital literacy have not yet been fully realized. Online social capital can partly explain this connection. Both instrumental support and institutional support play important moderating roles. The analysis also reveals that farmers who prefer cooperative entrepreneurship and those located in areas with well-developed rural logistics are more significantly influenced by digital literacy.

Keywords

e-commerce entrepreneurship intention-behavior gap digital literacy online social capital organizational support

Introduction

E-commerce entrepreneurship is considered a promising way to generate income and employment for the poorer strata of society (Zhao et al., 2023). According to the World Bank’s Digital Progress and Trends Report 2023, improvements in digital infrastructure, including localized logistics systems, local currency payment options, and multilingual interface development, have significantly influenced consumer purchase intentions and participation, which in turn has promoted the development of e-commerce entrepreneurship (The World Bank, 2024). This phenomenon is particularly evident in the agricultural economy. In China, the rural e-commerce has shown significant growth. According to data from the Ministry of Commerce of China, in 2024, online retail sales in rural areas and for agricultural products increased year-on-year by 6.4% and 15.8% respectively, marking an acceleration in the digital transformation of China’s agricultural economy. Farmers are key actors at the upstream end of the rural e-commerce industry chain. If farmers can become genuine e-commerce entrepreneurs, it would not only help expand the marketing channels for agricultural products but also serve as an effective measure for achieving high-quality development of the agricultural economy (Haji, 2021). However, although market opportunities have stimulated farmers’ intentions to engage in e-commerce entrepreneurship, many of them find it difficult to turn their intentions into actual entrepreneurial actions.

Intention-Behavior Gap: The Dilemma Under Multiple Constraints

The dilemma of farmers having the “intention but no action” in e-commerce entrepreneurship arises mainly from both macro- and micro-level factors. At the macro level, systemic barriers stem from factor market failures, a low degree of organizational development, and lagging institutional support. At the micro level, endogenous constraints include path dependence on traditional farming practices, a lack of digital skills application, and weak social networks (Mei et al., 2020; Tang & Zhu, 2020). The essence of this dual dilemma lies in the institutional friction, capability gap, and organizational disconnection between traditional agricultural systems and the digital economy. Consequently, addressing these challenges necessitates cultivating farmers’ digital competencies, revitalizing social capital, and deploying context-sensitive entrepreneurship policies.

Digital Literacy: Empowerment Potential

Enhancing farmers’ digital literacy provides a breakthrough in narrowing the gap between their intentions to engage in e-commerce entrepreneurship and their actual behaviors. Scholars have begun to focus on the role of digital literacy, recognizing it as a necessary condition for farmers’ participation in digital rural development (F. Li et al., 2023; Prasandha & Susanti, 2022). In the context of e-commerce entrepreneurship in particular, a high level of digital and technological literacy has become a key factor in overcoming resource constraints (Chen et al., 2024; Fahmi et al., 2023). However, existing studies tend to examine the impact of digital literacy on either entrepreneurial intention or actual behavior in isolation, focusing on issues such as the mechanisms that stimulate farmers’ entrepreneurial intentions or the spatial spillover effects of entrepreneurial behavior. Although some scholars have mentioned the dilemma of the “intention–behavior” gap in farmers’ e-commerce entrepreneurship (Mei et al., 2020; Tang & Zhu, 2020), there remains a lack of in-depth analysis regarding the theoretical logic and empirical validation of the “intention without action” mismatch. This limits our understanding of how digital literacy empowers the transformation from intention to action in farmers’ e-commerce entrepreneurship and hinders the precision of policy interventions. Therefore, exploring how digital literacy influences the gap between farmers’ e-commerce entrepreneurial intentions and behaviors is an important topic of urgent theoretical and practical significance.

The Present Study: Delving Deeper

In view of this, our study incorporates digital literacy, online social capital, and organizational support into the theoretical framework based on empowerment theory, and empirically examines the relationship between digital literacy and the gap between Chinese farmers’ e-commerce entrepreneurial intention and behavior using micro-level survey data. This study aims to answer the following questions:

RQ1: What is the relationship between digital literacy and the intention-behavior gap in farmers’ e-commerce entrepreneurship?

RQ2: How does online social capital influence the relationship between digital literacy and the intention-behavior gap? Does organizational support also play an important role?

RQ3: How are farmers’ entrepreneurial preferences and regional logistics infrastructure associated with the relationship between digital literacy and the intention-behavior gap?

The potential contributions of our study are as follows: First, we break through the traditional dichotomy of “intention–behavior” in the field of rural entrepreneurship by constructing an explanatory framework for the intention–behavior gap in farmers’ e-commerce entrepreneurship, thereby expanding the research boundaries of this field. Second, we reveal the black-box mechanisms and contingent conditions underlying the impact of digital literacy on the intention–behavior mismatch. From the perspective of resource empowerment within empowerment theory, the study empirically tests the transmission effect of online social capital in the process of digital empowerment and identifies the differentiated moderating effects of various forms of organizational support. This provides a scientific basis for developing a targeted empowerment policy system that integrates “individual capabilities and organizational support.” Third, our findings offer important insights for optimizing strategies to improve farmers’ digital literacy, refining e-commerce entrepreneurship policies, and providing useful references for promoting high-quality agricultural development.

The rest of this paper is arranged as follows: Section 2 reviews the relevant literature and proposes hypotheses. Section 3 introduces the data sources, variable selection, and research methodology. Section 4 shows the research results, and Section 5 does the analysis and discussion. Section 6 provides the research conclusions, theoretical and practical implications, as well as directions for future research.

Literature Review and Hypotheses Development

Literature Review

The occurrence of farmers’ e-commerce entrepreneurial behavior is the result of the combined effects of individual choices and external environments. In terms of individual choices, existing research indicates that factors such as technological capabilities, family resource endowments, and social networks significantly impact farmers’ entrepreneurial decisions (Liu et al., 2021; Song et al., 2024). Regarding the external environment, policy initiatives have effectively facilitated the transition of traditional farmers into modern operators (Cui et al., 2017; Haji, 2021). Additionally, some scholars have analyzed the aggregation phenomena observed in certain typical Taobao Villages, Taobao Towns, and rural e-commerce industrial parks in China (Mei et al., 2020; Qi et al., 2019).

However, when farmers attempt to translate their entrepreneurial intentions into actual actions, they still face multiple constraints. In terms of business models, existing research indicates that most farmers currently rely on decentralized operations and have yet to establish effective cooperation and mutual assistance mechanisms (Tang & Zhu, 2020). Regarding market competition, homogeneous competition and industrial homogenization are relatively common. The simple replication of entrepreneurial models has created developmental bottlenecks and dampened farmers’ enthusiasm for entrepreneurship. Furthermore, with the iterative upgrading of digital technologies, technological barriers and insufficient capabilities have also severely hindered the transformation of farmers’ e-commerce entrepreneurial behavior (Mei et al., 2020; Xie et al., 2022).

Scholars have begun to focus on the positive impact of digital literacy on farmers’ production and operational behaviors, considering it a core element for farmers’ participation in digital rural practices. Digital literacy helps farmers better adapt to the demands of the digital era and provides new opportunities and pathways for entrepreneurship (F. Li et al., 2023; Prasandha & Susanti, 2022). In the field of e-commerce, a high level of digital literacy has become key to overcoming resource constraints, as it mitigates knowledge gaps and skill deficiencies, thereby promoting farmers’ participation in e-commerce or engagement in e-commerce entrepreneurship. Empirical studies based on survey data further indicate that digital literacy significantly enhances both farmers’ willingness to participate in e-commerce and their level of involvement (Chen et al., 2024; Fahmi et al., 2023).

In summary, existing research has laid a solid foundation for this study, yet the following aspects require further exploration: Firstly, current literature has paid insufficient attention to the phenomenon of “intention without action” in farmers’ e-commerce entrepreneurship and has not systematically elucidated the potential relationship between digital literacy and the intention-behavior gap in this context. Secondly, the relationship between digital literacy and the intention-behavior gap, along with the underlying mechanisms, remains inadequately investigated. Thirdly, the boundary conditions and specific contexts of this relationship need further clarification. This study focuses on the practical issue of the “intention-behavior gap in farmers’ e-commerce entrepreneurship.” Based on empowerment theory, it incorporates digital literacy, online social capital, and organizational support into an integrated analytical framework. A theoretical model is constructed to examine the relationship between digital literacy and the intention-behavior gap, and micro-level survey data is utilized to empirically analyze the associated effects and potential pathways.

Hypotheses Development

The Influence of Digital Literacy on the Intention-Behavior Gap in Farmers’ E-Commerce Entrepreneurship

Gilster (1997) defined “digital literacy” as an individual’s comprehensive ability to access, understand, and organize digital information. As research has progressed, the concept of digital literacy has continuously expanded to include not only the technical skills required to use digital tools, but also cognitive abilities and attitudes (Havrilova & Topolnik, 2017; C. Li et al., 2025). At the same time, major public organizations have proposed various digital literacy frameworks, among which UNESCO’s seven-dimension Global Framework of Digital Literacy is the most representative and widely recognized as a mature model. Therefore, digital literacy constitutes a comprehensive competency system. This study conceptualizes digital literacy as comprising media and operational literacy, information and data literacy, communication and collaboration literacy, content creation literacy, digital safety literacy, problem-solving literacy, and job-related literacy.

The theory of implementation intentions posits that forming implementation intentions is key to translating entrepreneurial intentions into actual behavior (Gollwitzer, 1999). The empowerment logic of digital literacy aligns well with the formation mechanism of implementation intentions. From the perspective of its specific empowering effects, digital literacy has a positive impact on the transformation of farmers’ e-commerce entrepreneurial intentions into actual behaviors, thereby effectively bridging the intention-behavior gap. According to empowerment theory (Zimmerman, 2000), digital literacy facilitates the transformation of farmers’ e-commerce entrepreneurial intentions into actual behaviors in stages through skill support, subjective self-drive, and co-growth mechanisms. First, skill support is the foundational component of digital literacy empowerment. By enabling the operation of digital tools and integration of information, it helps farmers overcome technical barriers during the entrepreneurial planning stage, transforming abstract intentions into actionable strategies. Second, the subjective self-drive mechanism triggers cognitive restructuring and motivational iteration through knowledge acquisition, enabling farmers to break free from traditional agricultural mindsets during the goal-setting stage, reassess entrepreneurial feasibility, and enhance the stability of the entrepreneurial process. Third, digital literacy empowers farmers to form an interactive cycle of technological mastery, cognitive iteration, and income growth. Technological mastery enhances the effectiveness of digital tool use; cognitive iteration promotes innovation in market strategies; and through engaging in e-commerce practices, farmers gradually realize economic returns. This interactive feedback loop further reinforces technological investment and cognitive upgrading, forming a virtuous cycle that strengthens entrepreneurial execution resilience (Hartmann et al., 2022).

In summary, the essence of the intention-behavior gap in farmers’ e-commerce entrepreneurship lies in a failure of transformation. According to empowerment theory, digital literacy directly enhances their entrepreneurial execution capacity through skill support, strengthens entrepreneurial identity and motivation through subjective self-driving force, and fosters sustained positive feedback through interactive cycles—collectively mitigating the intention-behavior gap. Based on the above theoretical reasoning, this study expects digital literacy to exert a significant negative influence on the intention-behavior gap. Therefore, we propose the following hypotheses:

H1: Farmers’ digital literacy has a negative impact on the intention-behavior gap, meaning that improving farmers’ digital literacy can effectively mitigate the gap between their willingness to engage in e-commerce entrepreneurship and their actual behavior.

The Influence of Digital Literacy on the Online Social Capital

Based on empowerment theory, online social capital represents a key manifestation of resource-based empowerment (Leong et al., 2015). Digital literacy has a clear effect on the accumulation of social capital. Empowerment theory emphasizes that enhancing individual capabilities promotes greater participation in social activities (Zimmerman, 2000). The improvement of farmers’ digital literacy transforms traditional social interaction patterns, facilitating the formation and accumulation of online social capital. In digital environments, online social capital can be categorized into bonding social capital and bridging social capital (Williams, 2006). Bonding social capital arises from “strong ties” characterized by similar identities and high levels of trust, whereas bridging social capital stems from “weak ties” marked by heterogeneity and loose connections.

Digital literacy contributes to online social capital through multiple mechanisms. First, it enables farmers to use digital tools to maintain frequent interaction within existing social networks, thereby reinforcing bonding social capital. At the same time, digital literacy helps transcend geographical boundaries and facilitates connections with more diverse groups, promoting the development of bridging social capital. Second, digital literacy fosters cross-domain collaboration. Through platforms such as social media and live-streaming e-commerce, farmers engage in broader social interactions. These interactions not only strengthen bonding social capital within strong-tie networks but also promote the construction of diverse bridging social capital through cross-regional and cross-sector partnerships (Park et al., 2013). Third, digital technologies enhance trust through mechanisms of security and transparency. By enabling traceable transaction histories and credit ratings, digital tools translate the moral constraints of traditional acquaintance-based societies into quantifiable and verifiable systems of trust. This ensures transactional integrity and promotes broader and safer cooperation. Therefore, we propose the following hypotheses:

H2: The digital literacy of farmers has a positive impact on online social capital.

The Influence of Online Social Capital on the Intention-Behavior Gap in Farmers’ E-commerce Entrepreneurship

Online social capital plays a significant role in narrowing the gap between farmers’ e-commerce entrepreneurial intentions and actual behaviors. It improves information flow and decision-making, especially in overcoming the market information blind spots caused by geographic isolation. The diverse information channels embedded in online social capital help transform entrepreneurial impulses into feasible decisions. Furthermore, online social capital enables resource integration and environmental adaptation. In particular, weak-tie networks within bridging social capital are instrumental in mobilizing resources such as logistics and customer bases. Digital tools allow for dynamic matching of these resources, increasing the operability of entrepreneurial plans. Finally, online social capital facilitates trust transmission and risk mitigation. Bonding social capital reduces the risk of early-stage collaboration through reputation-based constraints within homogeneous groups. Meanwhile, bridging social capital ensures transactional safety and transparency via platform-based credit systems and payment guarantees, thereby enhancing the reliability of entrepreneurial actions (Wang et al., 2025). Therefore, we propose the following hypotheses:

H3: The online social capital of farmers contributes to bridging the gap between their e-commerce entrepreneurial intentions and actual behaviors.

The Moderating Role of Organizational Support

Organizational support theory posits that organizations stimulate positive behaviors and foster mutually beneficial relationships by demonstrating care, recognition, and assistance toward their members (Eisenberger et al., 1990). Such support functions as both a motivational mechanism and a resource compensation strategy, helping individuals overcome challenges and increasing the likelihood of task success. Some scholars categorize organizational support into instrumental support and emotional support. Given that farmers’ e-commerce entrepreneurship is closely linked to the institutional environment (Liang et al., 2023), organizations also help address challenges such as weak bargaining power and limited sales channels through institutional means. This study divides organizational support into instrumental support, institutional support, and emotional support. Instrumental support refers to resources such as technical training and infrastructure. Institutional support ensures the sustainability of entrepreneurship through rules, contracts, and standardized operations. Emotional support is reflected in the trust, encouragement, and care provided by the organization, which enhances farmers’ sense of security and belonging.

First, we discuss the moderating role of instrumental support. The effect of digital literacy on the gap between farmers’ e-commerce entrepreneurial intentions and actual behaviors may vary depending on the level of instrumental support. When instrumental support is high, organizations provide essential production resources, technical training, and equipment (Rhoades & Eisenberger, 2002), which significantly enhance farmer s’ e-commerce capabilities and reduce technological barriers. In contrast, when instrumental support is low, farmers may have entrepreneurial intentions but struggle to act due to a lack of resources. Moreover, the ability of digital literacy to reduce the intention-behavior gap through online social capital also differs depending on instrumental support level. With high instrumental support, farmers are better able to leverage online social capital to build commercial networks (Cofré-Bravo et al., 2019), thereby strengthening the mediating role of online social capital. In this context, the effect of digital literacy on entrepreneurial behavior through online social capital becomes more pronounced. Conversely, under conditions of low instrumental support, limitations in resources and tools hinder farmers from fully utilizing online social capital, weakening their ability to integrate resources and constraining the mediating effect of online social capital. Therefore, we propose the following hypotheses:

H4a: Instrumental support moderates the direct effect of digital literacy on the gap between farmers’ e-commerce entrepreneurial intentions and actual behaviors.

H4b: Instrumental support moderates the relationship between online social capital and the gap between e-commerce entrepreneurial intention and behavior.

Second, we discuss the moderating role of institutional support. The impact of digital literacy on the gap between farmers’ e-commerce entrepreneurial intentions and actual behaviors may vary depending on the level of institutional support. When institutional support is strong, contractual agreements and production standards help farmers stabilize their supply chains (Zhao et al., 2023), reducing market entry risks. In such cases, digital literacy and institutional support work synergistically to further facilitate the transformation of entrepreneurial intentions into actual behavior. In contrast, when institutional support is weak, the absence of clear rules and insufficient contractual protection leads to supply chain uncertainty, weakens market competitiveness, and hinders the realization of entrepreneurial behavior. Moreover, the effect of digital literacy in reducing the intention-behavior gap through online social capital also varies with the level of institutional support. When institutional support is strong, rules and contract enforcement enhance the trust foundation of online social capital, enabling farmers to effectively integrate resources and convert them into entrepreneurial actions (Mukul & Sheeri, 2024). Conversely, when institutional support is weak, problems such as information asymmetry, contract breaches, and weak brand influence constrain the effectiveness of online social capital, thereby weakening its mediating role. Therefore, we propose the following hypotheses:

H5a: Institutional support moderates the direct effect of digital literacy on the gap between farmers’ e-commerce entrepreneurial intentions and actual behaviors.

H5b: Institutional support moderates the relationship between online social capital and the gap between e-commerce entrepreneurial intention and behavior.

Third, we discuss the moderating role of emotional support. The impact of digital literacy on the gap between farmers’ e-commerce entrepreneurial intentions and actual behaviors may be influenced by differences in emotional support. When emotional support is high, farmers receive sufficient trust and guidance (George et al., 2021), which reduces entrepreneurial uncertainty and enhances the role of digital literacy in promoting e-commerce entrepreneurial behavior. In contrast, when emotional support is low, a lack of trust and encouragement limits farmers’ entrepreneurial motivation, thereby weakening the effect of digital literacy. Additionally, the effectiveness of digital literacy in narrowing the intention–behavior gap through online social capital is also influenced by emotional support. Under high emotional support, trusted networks strengthen the resource integration function of online social capital, enhancing the impact of digital literacy on entrepreneurial behavior. However, when emotional support is low, trust networks are fragile (Lyon, 2000), making it difficult for online social capital to function effectively. Therefore, we propose the following hypotheses:

H6a: Emotional support moderates the direct effect of digital literacy on the gap between farmers’ e-commerce entrepreneurial intentions and actual behaviors.

H6b: Emotional support moderates the relationship between online social capital and the gap between e-commerce entrepreneurial intention and behavior.

In summary, this study follows the research logic of “empowerment–process–outcome,” positioning digital literacy as the empowering driver of farmers’ e-commerce entrepreneurship, with online social capital and organizational support as key process variables. The gap between farmers’ e-commerce entrepreneurial intentions and behaviors is treated as the outcome, aiming to deeply examine how digital literacy influences the transformation of entrepreneurial intentions into actual behavior. Figure 1 illustrates the hypothesis model of this study.

Figure 1.

Theoretical hypotheses framework.

Methodology

Research Design and Data Collection

This study collected data from farmers through a questionnaire survey. All participants were informed about the study’s purpose, the voluntary nature of their participation, and their right to withdraw. The research design effectively minimized risks through strict anonymity and the non-collection of personally sensitive information. The survey consisted of two stages: the first stage involved a pilot survey conducted in August 2024, during which 65 farmers were randomly interviewed, and the questionnaire design was refined based on their feedback. The second stage, the formal survey, took place from November to December 2024. In this stage, a combination of stratified sampling and random sampling was used. One province was selected from each of China’s eastern, central, western, and northeastern regions. Two cities were chosen from each province, and 2 to 3 counties were selected from each city. Then, 2 to 5 townships were randomly selected from these counties, and finally, 10 to 20 farmers were randomly chosen from each township to complete the questionnaire survey. A total of 1,453 questionnaires were distributed, with 1,350 valid responses collected, resulting in a valid response rate of 92.91%. The questionnaire covered five key areas: basic household information, digital literacy, online social capital, entrepreneurial intention and decision-making, and organizational support. The valid sample of farmers had the following characteristics: male household heads accounted for approximately 71.85% of the sample; most respondents were older, with 62.67% aged 50 or above; education levels were generally low, with 75.26% having a high school education or below, including vocational and technical schooling; and approximately 60.15% of the households had two members in the labor force.

Measures

Dependent Variable

The intention-behavior gap (IBG) in farmers’ e-commerce entrepreneurship refers to the discrepancy between entrepreneurial intention and actual behavior. This study focuses on the misalignment between farmers’ intention to engage in e-commerce entrepreneurship and their real-world entrepreneurial actions, aiming to explore the internal transformation mechanism from “having intention but no action” to “having both intention and action.” Accordingly, after excluding samples without e-commerce entrepreneurial intention, we coded farmers who had actually engaged in e-commerce entrepreneurship as 0, indicating no intention-behavior gap, and those who had not engaged as 1, indicating the presence of an intention-behavior gap. After removing the samples lacking entrepreneurial intention, a total of 952 farmers were identified as having the intention to engage in e-commerce entrepreneurship, accounting for 70.52% of the total sample. However, only 312 of them had actually carried out e-commerce entrepreneurial activities, while the remaining 640 farmers failed to convert their intention into action, accounting for 67.23% of the farmers with entrepreneurial intention. This further confirms the existence of the widespread “having intention but no action ” phenomenon in China.

Independent Variables

The key independent variable is Digital Literacy (DL). We referenced the Digital Literacy Global Framework released by UNESCO in 2018 (Law et al., 2018), which is applicable globally, and adapted it to the digital environment of Chinese farmers. Based on this, we set up 23 items across seven dimensions, which include media literacy, information and data literacy, social literacy, content creation literacy, digital security literacy, problem-solving literacy, and agriculture-related literacy. The respondents were asked to rate each item on a five-point Likert scale, ranging from 1 to 5, where 1 stands for “totally disagree” and 5 for “totally agree.” These have been discussed in Table 1.

Table 1.

The Evaluation Index System for Farmers’ Digital Literacy.

Item	Description
Media literacy (DL1)	I know how to connect to a wireless network (WiFi).
	I know how to install and uninstall software (apps) on a mobile phone.
	I know how to use a mobile phone and other devices to take photos and record videos.
information and data literacy (DL2)	I know how to use different browsers to search for information.
	I know how to bookmark a webpage in the browser.
	I know how to access previously visited webpages.
	I know how to access a website by clicking a link.
Social literacy (DL3)	I know how to use social apps like WeChat, Douyin, and others to share videos and pictures.
	I know how to use social apps like WeChat, Douyin, and others to communicate with others.
	I know how to remove people I don’t communicate with from the friend list on social apps like WeChat and Douyin.
Content creation literacy (DL4)	I know how to use software to publish my opinions online.
	I know how to edit videos.
	I know how to edit images.
Digital security literacy (DL5)	I know how to avoid being subjected to online harassment.
	I know how to protect my personal privacy and security online.
	I know how to identify which content is an advertisement.
Problem-solving literacy (DL6)	I know how to seek help from others when encountering problems with a new device.
	I know that there are some things that should not be casually shared with strangers online.
	I am aware of the limitations in my use of internet technologies.
Job-related literacy (DL7)	I know some specialized digital tools and technologies in the field of agriculture.
	I know how to use digital technology to improve agricultural production and processing efficiency.
	I know how to use digital technology to learn agricultural-related skills and knowledge.
	I know how to use online platforms to sell products and increase sales.

Mediating Variables

The mediating variable, online social capital (OSC), was measured based on the Online Social Capital Scale proposed by Williams (2006), and was assessed from two dimensions: bridging social capital and bonding social capital. Specifically, bridging social capital is measured using five items: “I often communicate with members I am not familiar with in WeChat groups,”“I am very interested in the opinions expressed by different people in WeChat groups,”“I often discuss planting techniques with unfamiliar members in WeChat groups,”“I can always meet new people through WeChat groups,” and “I am willing to exchange sales information with unfamiliar members in WeChat groups.” Bonding social capital is measured with five items: “When making important decisions, I can seek advice from family and friends through online social platforms and e-commerce platforms,”“My family and friends often share planting and sales information with me through online social platforms and e-commerce platforms,”“My family and friends frequently recommend distributors to me through online social platforms and e-commerce platforms,”“I can borrow money from family and friends through online social platforms,” and “I often share my personal life with family and friends on online social platforms.” The respondents were asked to rate each item on a five-point Likert scale, ranging from 1 to 5, where 1 stands for “totally disagree” and 5 for “totally agree.”

Moderate Variables

The moderate variable, Organizational Support (OS), is based on the Organizational Support Scale developed by Eisenberger et al. (1990), and adapted to the context of e-commerce entrepreneurship among Chinese farmers. It is measured using 13 items across three dimensions: Instrumental support (ITS), Institutional Support (INS), and Emotional Support (ES). Instrumental support is measured by items such as “The organization provides me with the necessary element services for agricultural production and product processing,”“The organization provides me with technical services, such as technical personnel and guidance,”“The organization provides me with financial services, such as cash support and low-interest loans,”“The organization provides me with informational services, such as policy interpretation and market outlook,” and “The organization provides me with training services, such as entrepreneurial training and organizational visits.” Institutional Support is represented by items like “The organization signs contracts or product orders with me,”“The organization sets standards for product production and processing,”“The organization provides me with brand-building services,” and “The organization provides me with product sales services.” Emotional Support is measured by items such as “The organization respects my decisions in agricultural production,”“The organization trusts me greatly,”“When encountering problems, the organization provides me with guidance and support,” and “ The organization cares about my well-being in agricultural production.” The respondents were asked to rate each item on a five-point Likert scale, ranging from 1 to 5, where 1 stands for “totally disagree” and 5 for “totally agree.”

Control Variables

We comprehensively considered both internal characteristics and external factors that influence farmers’ e-commerce entrepreneurship intentions and behaviors. Nine control variables were selected from three aspects. The first aspect is the individual characteristics of farmers, including gender, age, education level, health status, and entrepreneurial channel preference. The second aspect is family characteristics, such as the number of laborers and average annual family income. The third aspect is the external environment, which includes geographical location and logistics conditions.

Variables are described in Table 2.

Table 2.

Descriptive Statistics of Variables.

Variable	Variable definition
Dependent variable
IBG	Intention and behavior gap: Having intention but no behavior = 1, Having both intention and behavior = 0
Core independent variable
DL	Digital literacy: Calculation results of Section 3.3.2
Mediating variable
OSC	Online social capital: Calculation results of Section 3.3.3
Moderate variable
ITS	Instrumental support: Calculation results of Section 3.3.4
INS	Institutional support: Calculation results of Section 3.3.4
ES	Emotional support: Calculation results of Section 3.3.4
Control variables
Gender	Male = 1, Female = 0
Age	The age of respondents
Edu	Primary school and below = 1, Junior middle school = 2, High school = 3, Junior college = 4, Bachelor’s degree or above = 5.
Health	Very unhealthy = 1, Somewhat unhealthy = 2, Average = 3, Somewhat healthy = 4, Very healthy = 5
ECP	Independent entrepreneurship preference = 1, Cooperative entrepreneurship preference = 0,
Labor	Number of household labor force
Income	Annual per capita income: Below 50,000 yuan = 1, 50,000 to 100,000 yuan = 2, 100,000 to 150,000 yuan = 3, 150,000 to 200,000 yuan = 4, 200,000 yuan and above = 5
Distance	Distance from your home to the nearest town: Within 2 km = 1, 2 to 5 km = 2, 5 to 10 km = 3, 10 to 15 km = 4, 15 km or above = 5
Logistics	Number of courier service points within 1 km of your residence: None = 1, 1 point = 2, 2 points = 3, 3 points = 4, 4 points or more = 5

Data Estimation Techniques

Baseline Regression Model

The intention-behavior gap in farmers’ e-commerce entrepreneurship is a binary discrete variable. Therefore, a Probit regression model is constructed to empirically analyze the impact of digital literacy on this intention-behavior gap.

Prob (Y = 1 | x_{i}) = Φ (β_{0} + k C A + \sum_{i = 1}^{9} β_{i} . X_{i} + λ_{FE} + ε_{i})

(1)

In Equation 1, Y represents the intention-behavior gap in farmers’ e-commerce entrepreneurship; CA represents digital literacy; X represents a series of control variables; $β_{0}$ , k and $β_{i}$ are the parameters to be estimated for each variable; $λ_{FE}$ represents regional fixed effects; $ε_{i}$ is the random error term.

It is worth noting that though we’ve mitigated potential endogeneity from omitted variables by controlling for individual, household, and external environmental factors, the baseline regression results may still be biased due to unobserved factors and reverse causality. As farmers’ e-commerce entrepreneurial intentions and behaviors align more, their digital literacy may naturally improve with frequent exposure to digital environments, intensifying reverse causality. To address this, we employ an IV-Probit model. This approach uses instrumental variables to break the correlation between the endogenous regressor (digital literacy) and the error term, thereby enabling the identification of its causal effect on entrepreneurial intentions and behaviors and correcting for bias from reverse causality.

Bootstrap Approach

This study uses the Bootstrap method in the PROCESS macro to test the mediating effect of online social capital and the moderating effect of organizational support, mainly for the following reasons: (1) Bootstrap relaxes the normality assumption by constructing an empirical distribution through repeated resampling, replacing parametric assumptions. This significantly enhances the robustness of confidence interval estimation for non-normally distributed data or small samples. (2) It is applicable to analyzing mediation and moderation effects with binary dependent variables. (3) Compared with traditional mediation analysis methods, Bootstrap estimates the path coefficients of the full model and constructs bias-corrected confidence intervals, effectively reducing the risk of suppression effects and Type II errors. (4) By testing moderated mediation effects within a unified framework, Bootstrap can control for covariate interference across multiple paths and reduce omitted variable bias from stepwise modeling (Hayes, 2017).

Data Analysis and Results

Common Method Bias

We used Harman’s single factor test to examine common method bias (Podsakoff et al., 2003). The results showed that the variance explained by the first unrotated principal component was 25.16%, indicating that there was no serious common method bias in our measurement process.

Reliability and Validity Tests

Table 3 shows the reliability and validity test results. The Cronbach’s α coefficients and CR values for digital literacy, online social capital, and organizational support all exceed 0.9, surpassing the commonly accepted threshold of 0.7, indicating high internal consistency (Hair et al., 2009). The KMO values are above 0.7 and the p-values of Bartlett’s test of sphericity are less than 0.01, suggesting the data are suitable for factor analysis. Also, the cumulative variance explained by each variable is over 70%, well above the 40% benchmark, showing strong explanatory power of the measurement items. For validity, all standardized factor loadings are above 0.6, and the AVE values exceed 0.5, demonstrating high convergent validity. Moreover, the square roots of the AVEs for each construct are greater than the inter-construct correlation coefficients, indicating good discriminant validity (Fornell & Larcker, 1981). In short, the scale passes the reliability and validity tests.

Table 3.

Reliability and Validity Test of Variables.

						Bartlett’s test of sphericity
Variable	Cronbach’s α	Factor Load	Cumulative variance contribution rate	CR	AVE	KMO	p-Value
DL	.964	0.701–0.797	88.581%	0.965	0.550	0.950	.000
OSC	.952	0.618–0.931	82.771%	0.948	0.650	0.936	.000
TS	.954	0.831–0.843	88.641%	0.952	0.604	0.937	.000
IS	.970	0.788–0.802
ES	.968	0.687–0.725

Hypothesis Testing of the Main Effects

Baseline Regression

The baseline regression results are in Table 4. Column (1) shows digital literacy negatively impacts IBG without control variables. In Column (2), after adding control variables, the negative impact of digital literacy on IBG remains significant at the 1% level, supporting H1. Column (3) presents a further analysis of how different dimensions of digital literacy affect IBG. Among the seven dimensions, media operation and occupation-related literacy have significant negative effects at the 5% level, while information and data literacy and social literacy show significant negative effects at the 1% level. The other dimensions have no statistically significant effects. These findings suggest that at present, the key digital literacy aspects helping reduce farmers’ inconsistency between entrepreneurial intention and behavior in e-commerce mainly lie in basic digital competencies. As digital technologies advance, the potential of digital literacy to promote intention-behavior consistency among farmers needs further exploration. Specifically, improving higher-level digital literacies like digital content creation, cybersecurity awareness and skills, and complex problem-solving ability is expected to drive farmers’ entrepreneurial decisions in e-commerce.

Table 4.

Regression Results of the Influence of Farmers’ Digital Literacy on IBG.

	Probit
	IBG	IBG	IBG
Variable	(1)	(2)	(3)
DL	−0.690*** (0.114)	−0.709*** (0.119)
DL1			−0.096** (0.044)
DL2			−0.190*** (0.044)
DL3			−0.166*** (0.047)
DL4			−0.039 (0.044)
DL5			−0.058 (0.045)
DL6			−0.068 (0.044)
DL7			−0.085** (0.043)
Gender		−0.060 (0.102)	−0.065 (0.102)
Age		0.018*** (0.004)	0.017*** (0.004)
Edu		−0.014 (0.057)	−0.024 (0.057)
Health		−0.009 (0.047)	−0.009 (0.048)
ECP		−0.483** (0.167)	−0.509** (0.166)
Labor		−0.008 (0.048)	−0.011 (0.048)
Income		−0.070 (0.115)	−0.079 (0.116)
Distance		0.071 (0.043)	0.071 (0.043)
Logistics		−0.219*** (0.057)	−0.223*** (0.057)
Region FE	Yes	Yes	Yes
Con_s	0.707*** (0.128)	0.407 (0.519)	0.514 (0.521)
N	952	952	952
Pseudo R²	.0453	.0995	.1064

indicates p < .05, *** indicates p < .01.

Endogeneity Test

This study tackles endogeneity issues by employing the IV-Probit. We choose an aggregation-level instrumental variable, specifically using “the average digital literacy level of other farmers in the same township, excluding the surveyed household head” as the instrumental variable (IV-DL). The rationale behind this instrumental variable is grounded in two key aspects. Firstly, farmers within the same township are exposed to similar digital infrastructures, training resources, and policy environments. This makes an individual farmer’s digital literacy vulnerable to the influence of the overall digital atmosphere in the area, thereby fulfilling the correlation requirement between the instrumental variable and the endogenous variable. Secondly, the average digital literacy level of other farmers in the township is improbable to directly impact the e-commerce entrepreneurial decisions of a particular farmer, satisfying the exogeneity requirement.

From the IV-Probit model results presented in Table 5, we can observe that the corresponding p-value for the Wald test of exogeneity is .0307, which is significant at the 5% level. Consequently, we reject the null hypothesis that digital literacy is an exogenous variable, suggesting the existence of endogeneity issues. The F-value in the first stage of the IV-Probit model is 65.89, indicating that the selected instrumental variable does not have a weak-instrument problem. The second-stage regression results show that after accounting for endogeneity, farmers’ digital literacy still narrows the gap between their e-commerce entrepreneurial intentions and behaviors ( $β = - 0.956$ , $p < 0.01$ ). Overall, the aforementioned tests demonstrate that the instrumental variables selected in this study are effective and further bolster the reliability of the conclusions drawn in this paper.

Table 5.

Endogeneity Test.

	IV-Probit
Variable	First stage	Second stage
IV-DL	0.930*** (0.029)
DL		−0.956*** (0.165)
Control variables	Yes	Yes
Region FE	Yes	Yes
F-value in the first stage	65.89
Wald test p-value	.0307
N	952	952

***

indicates p < .01.

Mediating Effect Test

We used the PROCESS macro to test the mediation effect. Based on the inclusion of all control variables, a bootstrap sampling procedure was conducted 5,000 times at a 95% confidence level. The results are shown in Table 6. The bootstrap test for mediation effects indicates that after incorporating online social capital, the direct effect of digital literacy on the inconsistency between farmers’ e-commerce entrepreneurial intention and behavior remains significant, $β = - 0.621$ , $Boot 95 % CI = [- 1.077, - 0.164]$ , with the confidence interval excluding zero. The mediating effect of online social capital is also significant, $β = - 0.794$ , $Boot 95 % CI = [- 1.098, - 0.566]$ , with the confidence interval excluding zero. Therefore, after including the mediating variable, the direct negative effect of digital literacy on the inconsistency remains significant. Online social capital plays a partial mediating role, thus supporting H2 and H3.

Table 6.

Mediating Effect Report.

Effect	β	Boot SE	Boot LLCI	Boot ULCI
Direct effect	−0.621	0.233	−1.077	−0.164
Indirect effect	−0.794	0.136	−1.098	−0.566

Moderating Effect Test

We used the PROCESS macro to test the moderating effects of the three dimensions of organizational support. Results are shown in Tables 7 and 8. The interaction term between digital literacy and instrumental support negatively affects IBG, $β = - 0.601$ , $Boot 95 % CI = [- 1.245, - 0.082]$ , with a confidence interval excluding zero. The interaction term between online social capital and instrumental support also significantly negatively affects IBG, $β = - 0.449$ , $Boot 95 % CI = [- 0.766, - 0.160]$ , indicating that instrumental support moderates the relationship between online social capital and IBG. The index of moderated mediation in Table 8 is −0.224, $Boot 95 % CI = [- 0.404, - 0.077]$ , with a confidence interval excluding zero. Thus, H4a and H4b are supported.

Table 7.

Moderating Effect Report.

				95% confidence interval
Model	Variable	β	Boot SE	Boot LLCI	Boot ULCI
Model 1 (Moderate variable: ITS)	DL	−0.539	0.283	−1.112	0.018
	OSC	−1.610	0.150	−1.984	−1.382
	TS	−0.009	0.098	−0.207	0.180
	DL×ITS	−0.601	0.298	−1.245	−0.082
	OSC×ITS	−0.449	0.153	−0.766	−0.160
	Con_s	1.753	1.057	−0.213	3.933
	Control variables	Yes	Yes	Yes	Yes
	Region FE	Yes	Yes	Yes	Yes
Model 2 (Moderate variable: INS)	DL	−0.675	0.281	−1.245	−0.135
	OSC	−1.619	0.152	−1.987	−1.391
	TS	−0.140	0.098	−0.333	0.054
	DL×INS	−0.653	0.296	−1.257	−0.104
	OSC×INS	−0.415	0.155	−0.760	−0.148
	Con_s	1.287	1.092	−0.765	3.514
	Control variables	Yes	Yes	Yes	Yes
	Region FE	Yes	Yes	Yes	Yes
Model 3 (Moderate variable: ES)	DL	−0.607	0.275	−1.167	−0.076
	OSC	−1.584	0.148	−1.939	−1.362
	TS	−0.058	0.100	−0.256	0.142
	DL×ES	−0.032	0.270	−0.551	0.512
	OSC×ES	−0.121	0.152	−0.425	0.176
	Con_s	0.967	1.032	−1.022	3.045
	Control variables	Yes	Yes	Yes	Yes
	Region FE	Yes	Yes	Yes	Yes

Table 8.

Index of Moderated Mediation.

Moderate variable	IMM	Boot SE	95% confidence interval
Moderate variable	IMM	Boot SE	Boot LLCI	Boot ULCI
ITS	−0.224	0.083	−0.404	−0.077
INS	−0.207	0.085	−0.397	−0.071
ES	−0.060	0.076	−0.217	0.086

Similarly, the interaction term between digital literacy and institutional support has a negative effect on IBG, $β = - 0.653$ , $Boot 95 % CI = [- 1.257, - 0.104]$ , with the confidence interval excluding zero. In addition, the interaction term between online social capital and institutional support also has a significant negative effect on IBG, $β = - 0.415$ , $Boot 95 % CI = [- 0.760, - 0.148]$ , with the confidence interval excluding zero. As shown in Table 8, the index of moderated mediation is −0.207, $Boot 95 % CI = [- 0.397, - 0.071]$ , and the confidence interval does not include zero, indicating that institutional support moderates the relationship between online social capital and IBG. Therefore, H5a and H5b are supported.

However, the interaction term between digital literacy and emotional support has no significant effect on IBG, $β = - 0.032$ , $Boot 95 % CI = [- 0.551, 0.512]$ , as the confidence interval includes zero. Similarly, the interaction between online social capital and emotional support also shows no significant effect on IBG, $β = - 0.121$ , $Boot 95 % CI = [- 0.425, 0.176]$ , with the confidence interval including zero. As shown in Table 8, the index of moderated mediation is −0.060, $Boot 95 % CI = [- 0.217, 0.086]$ , and the confidence interval includes zero, indicating that emotional support does not significantly moderate the relationship between online social capital and IBG in farmers’ e-commerce entrepreneurship. Therefore, H6a and H6b are not supported. The underlying reason is that although emotional support can alleviate farmers’ entrepreneurial stress, stimulate motivation, and provide psychological comfort, the core challenges they face—such as technology, market competition, and funding—are practical issues. Emotional support primarily functions at the psychological level and is difficult to translate directly into effective solutions for real-world problems.

Robustness Test

First, we replace the Probit model with the Logit model to conduct a robustness test of the baseline regression. As shown in Column (1) and (2) of Table 9, after changing the econometric model, the negative effect of digital literacy on IBG remains significant at the 1% level. In terms of dimensions, media operation literacy, information and data literacy, social literacy, and job-related literacy all significantly negatively affect IBG, while the other dimensions do not show statistical significance. These results are consistent with the baseline regression results, indicating that the baseline model is robust. Next, considering the significant differences in physical function, cognitive ability, and knowledge level between farmers aged 70 and above and younger farmers, their digital literacy and online social capital may not significantly alleviate IBG. After excluding the sample of elderly farmers over 70 years old, we re-conducted the Probit regression analysis. As shown in Column (3) and (4) of Table 9, after restricting the sample, the effects of digital literacy and its dimensions on IBG are consistent with the conclusions from the full sample analysis, further verifying the robustness of the baseline model.

Table 9.

Robustness Test.

	Logit		Sample alteration
	IBG	IBG	IBG	IBG
Variable	(1)	(2)	(3)	(4)
DL	−1.187*** (0.201)		−0.669*** (0.123)
DL1		−0.166** (0.074)		−0.090* (0.046)
DL2		−0.324*** (0.075)		−0.199*** (0.047)
DL3		−0.283*** (0.078)		−0.156*** (0.048)
DL4		−0.064 (0.072)		−0.026 (0.045)
DL5		−0.093 (0.075)		−0.059 (0.046)
DL6		−0.112 (0.073)		−0.040 (0.046)
DL7		−0.144** (0.071)		−0.084* (0.044)
Control variables	Yes	Yes	Yes	Yes
Region FE	Yes	Yes	Yes	Yes
Con_s	0.707 (0.882)	0.861 (0.885)	0.439 (0.541)	0.568 (0.543)
N	952	952	853	853
Pseudo R²	.0997	.1070	.0831	.0914

Indicates p < .10, **indicates p < .05, ***indicates p < .01.

Moreover, we constructed a Generalized Structural Equation Model (GSEM), and Figure 2 displays the results of the mediating effect pathways from the GSEM model. We will first interpret the regression model results for the first pathway (the upper pathway in Figure 2). The path coefficient for the first half, which originates from an OLS regression model, reveals a significant positive impact of digital literacy on online social capital ( $β = 0.499$ , $p < 0.01$ ). The path coefficient for the second half comes from a Logit regression model, indicating a significant negative impact of online social capital on IBG ( $β = - 1.591$ , $p < 0.01$ ). The second pathway in Figure 2 illustrates the direct impact of digital literacy on IBG, with a significantly negative coefficient value ( $β = - 0.620$ , $p < 0.01$ ).

Figure 2.

Mediation effect paths based on GSEM.

We introduce the KHB model to calculate the mediating effect, with the results presented in Table 10. It is evident that online social capital plays a partial mediating role in the impact of digital literacy on IBG.

Table 10.

Total, Direct, and Indirect Effects of Digital Literacy on IBG.

Effect	Coefficient	Standard error	Z-value	p-Value	95% confidence interval
Total effect	1.415	0.232	6.10	.000	[0.960, 1.869]
Direct effect	0.620	0.233	2.66	.008	[0.164, 1.077]
Indirect effect	0.794	0.115	6.88	.000	[0.568, 1.020]

Next, based on the model in Figure 2, we incorporate three moderating variables. As shown in Model (1) and (2) of Table 11, the interaction terms of digital literacy and instrumental support, as well as online social capital and instrumental support, both have a significant impact on IBG, indicating that the moderating effect of instrumental support is significant. From Model (3) and (4) of Table 11, the interaction terms of digital literacy and institutional support, as well as online social capital and institutional support, both have a significant impact on IBG, indicating that the moderating effect of institutional support is significant. Model (5) and (6) of Table 11 examine the moderating effect of emotional support. The results show that although the coefficients of the interaction terms of digital literacy and emotional support, as well as online social capital and emotional support, are negative, they are not significant, indicating that the moderating effect of emotional support is not significant. The above conclusions are consistent with the previous findings.

Table 11.

Moderating Effect Paths Based on GSEM.

Path	(1)	(2)	(3)	(4)	(5)	(6)
DL→IBG	−0.519*	−0.495*	−0.662**	−0.558**	−0.610**	−0.610**
DL→OSC	0.499***	0.499***	0.499***	0.499***	0.499***	0.499***
OSC→IBG	−1.570***	−1.629***	−1.574***	−1.626***	−1.584***	−1.584***
ITS→IBG	−0.106	−0.007
DL×ITS→IBG	−0.760**
OSC×ITS→IBG		−0.500***
INS→IBG			−0.240***	−0.153
DL×INS→IBG			−0.747***
OSC×INS→IBG				−0.471***
ES→IBG					−0.080	−0.059
DL×ES→IBG					−0.049
OSC×ES→IBG						−0.122

Indicates p < .10, ** indicates p < .05, *** indicates p < .01.

Heterogeneity Analysis

Based on farmers’ preferences for e-commerce entrepreneurship channels, we divide farmers into cooperative entrepreneurship groups and independent entrepreneurship groups. The results in Table 12 show that the negative impact of digital literacy on IBG is mainly reflected in the group of farmers who prefer cooperative entrepreneurship. Cooperative entrepreneurship, relying on resource sharing, logistics support, and traffic diversion, can significantly enhance the effectiveness of digital literacy conversion. In contrast, farmers who prefer independent entrepreneurship can obtain technological and market information through digital literacy, but due to the lack of a stable cooperation mechanism, their entrepreneurial intentions are difficult to convert into actual behavior, limiting the positive role of digital literacy. In addition, based on whether there’s an express delivery station within 1 km of a farmer’s residence, we divide farmers into two groups: those with and without express delivery access. As shown in Table 12, the impact of digital literacy is mainly evident among farmers in areas with express delivery access. For farmers without nearby delivery services, logistical bottlenecks hinder timely and effective product delivery to consumers, reducing their entrepreneurial confidence. Consequently, digital literacy is less effective in bridging the gap between farmers’ e-commerce entrepreneurial intentions and actual behavior.

Table 12.

Heterogeneity Analysis.

	Farmers’ preference for entrepreneurial channels		Rural logistics situation
	Cooperative entrepreneurshipIBG	Independent entrepreneurshipIBG	With courier service pointsIBG	Without courier service pointsIBG
Variable	(1)	(2)	(3)	(4)
DL	−0.742*** (0.124)	−0.287 (0.502)	−0.766*** (0.145)	−0.272 (0.257)
Control variables	Yes	Yes	Yes	Yes
Region FE	Yes	Yes	Yes	Yes
Con_s	0.317	0.029	1.990	−0.325
Con_s	(0.550)	(2.029)	(0.645)	(1.126)
N	880	72	752	200
Pseudo R²	.1007	.2926	.1714	.0723

***

indicates p < .01.

Discussion

This study aims to explore the relationship between digital literacy and IBG in farmers’ e-commerce entrepreneurship, as well as its underlying mechanisms. H1 suggests that digital literacy has a negative impact on IBG. The results support this hypothesis ( $β = - 0.709$ , $p < 0.01$ ), indicating that as the level of digital literacy among farmers increases, the gap between their intention and actual behavior in e-commerce entrepreneurship significantly narrows. This finding is consistent with the viewpoints of existing research, which propose that digital capabilities can assist individuals in overcoming operational obstacles and cognitive limitations during the entrepreneurial process, thereby facilitating the execution of actions (Chen et al., 2024; Fahmi et al., 2023).

H2 and H3 suggest that online social capital mediates the relationship between digital literacy and IBG. The results support these hypotheses, with a 95% confidence interval for the indirect effect excluding zero ( $β = - 0.794$ , $Boot 95 % CI = [- 1.098, - 0.566]$ ). This indicates that digital literacy not only directly reduces the gap between intention and behavior but also indirectly promotes the implementation of e-commerce entrepreneurial behavior through online social capital. Our findings further reveal the social pathway through which digital literacy operates, suggesting that in the context of digital entrepreneurship, social capital remains a crucial vehicle for resource integration and behavioral transformation. Studies of Turkish rural e-commerce underscore online social capital as a fundamental resource for entrepreneurship, enabling the crucial connection between localized production and global market access (Yaşlak et al., 2023).

Furthermore, we introduced instrumental support and institutional support as moderating variables. The results support H4a ( $β = - 0.601$ , $Boot 95 % CI = [- 1.245, - 0.082]$ ) and H4b ( $β = - 0.449$ , $Boot 95 % CI = [- 0.766, - 0.160]$ ), indicating that instrumental support not only significantly moderates the direct impact of digital literacy on IBG, but also moderates the mediating pathway of online social capital. Similarly, H5a ( $β = - 0.653$ , $Boot 95 % CI = [- 1.257, - 0.104]$ ) and H5b ( $β = - 0.415$ , $Boot 95 % CI = [- 0.760, - 0.148]$ ) are also supported, demonstrating that institutional support exerts a significant moderating effect in both the direct pathway and the mediating pathway. This implies that when farmers have access to more technical support and institutional guarantees, the promoting effects of digital literacy and online social capital on behavioral realization become more pronounced. Our findings provide nuanced directions for policy formulation, suggesting that the overall efficacy of digital literacy development should be enhanced through the establishment of supporting tools and an enabling institutional environment. These findings align with the research by Rubyutsa et al. (2024), which noted that the “train-the-trainer” approach, as a form of institutional support, combined with instrumental support provided through weekly group meetings, collectively facilitated the success of rural entrepreneurship in Rwanda.

However, the moderating effect of emotional support did not receive empirical support, with both H6a ( $β = - 0.032$ , $Boot 95 % CI = [- 0.551, 0.512]$ and H6b ( $β = - 0.121$ , $Boot 95 % CI = [- 0.425, 0.176]$ ) being insignificant. Although emotional support is considered an important factor in certain contexts, in this study, its moderating effects on the direct impact of digital literacy and the mediating pathway of online social capital did not reach statistical significance. The development of rural e-commerce in China has entered a stage characterized by platform economy and resource-driven approaches, where entrepreneurial actions rely more on rationalized mechanisms such as technology, processes, and institutions. In behaviors like farmers’ e-commerce entrepreneurship, which emphasize practical operations and resource integration, instrumental support and institutional support possess more practical regulatory roles compared to emotional support. The research by Haji (2021) also points out that in the development of rural e-commerce in BRICS countries, although socio-cultural factors are important, substantive support such as infrastructure and education appears to be more crucial.

In addition, our heterogeneity results show that digital literacy’s impact depends on conditions. Cooperative entrepreneurship and accessible logistics enable its full role. Zeng et al. (2024) and Keikhay Farzane et al. (2024) also confirmed the importance of cooperative entrepreneurship and logistics conditions. This means that rural e-commerce policies should go beyond individual skill training and focus on the construction of cooperation mechanisms and the improvement of logistics.

Conclusions and Implications

Conclusions

This study was based on empowerment theory to construct a theoretical framework for the gap between farmers’ e-commerce entrepreneurial intention and behavior. We investigated the relationship between digital literacy and the intention-behavior gap among Chinese farmers, using mediation and moderation analyses to identify correlating factors. Our findings indicate that: First, there is a significant negative relationship between digital literacy and the gap between farmers’ e-commerce entrepreneurial intention and behavior. Foundational digital literacies, such as media operation, information and data processing, social literacy, and agriculture-related application capabilities, demonstrate a strong effect in facilitating the conversion of intention into actual behavior. Meanwhile, although advanced digital literacy holds potential value in theory, its practical efficacy has not yet been fully realized. Second, online social capital partially mediates the relationship between digital literacy and the intention-behavior gap. Instrumental support and institutional support not only strengthened the negative relationship between digital literacy and the intention-behavior gap, but also reinforced the mediating path of online social capital. In contrast, the moderating effect of emotional support was not statistically significant. Third, the heterogeneity analysis reveals that the negative relationship between digital literacy and the intention-behavior gap is more pronounced among farmers who prefer cooperative entrepreneurial models. Meanwhile, in regions with relatively well-developed rural logistics infrastructure, this association is also more prominently manifested.

Implications

Implications of Theory

The study makes theoretical contributions in the following three aspects: First, it constructs a theoretical analytical framework for the relationship between digital literacy and the intention-behavior gap in farmers’ e-commerce entrepreneurship. Most existing literature still examines the impact of digital literacy on either farmers’ entrepreneurial intention or behavior in isolation (Chen et al., 2024; Fahmi et al., 2023), with insufficient attention to the “ having intention but no action” phenomenon observed in farmers’ e-commerce entrepreneurial practices. Notably, with the rapid development of rural e-commerce in China, farmers have generally recognized the value of e-commerce entrepreneurship. However, in reality, there is a paradox where farmers “want to start a business” but “dare not” or “do not know how to” start a business. This paper reveals the intrinsic relationship between farmers’ digital literacy and the transformation of e-commerce entrepreneurial behavior, enriching the theoretical understanding of the inconsistency between entrepreneurial intention and entrepreneurial action. Second, existing literature has lacked in-depth exploration of the relationship between digital literacy and the conversion of farmers’ e-commerce entrepreneurial intention into behavior. This study finds that online social capital plays a mediating role. Third, this paper reveals the moderating effect of organizational support. Compared to urban entrepreneurs, farmers’ e-commerce entrepreneurial behavior is more dependent on organizational support provided by cooperatives, agricultural enterprises, and village collectives. However, existing literature has not identified the boundary conditions of organizational support, making it difficult to explain the differences in farmers’ entrepreneurial behavior transformation at the same level of digital literacy. By introducing the organizational support variable, this study provides a contextualized theoretical explanation for research on farmers’ e-commerce entrepreneurship.

Implications of Practice

Our findings suggest the following practical implications. First, it is essential to establish differentiated pathways for enhancing digital literacy, with a particular focus on consolidating foundational digital capabilities. This study proves that foundational digital operations, information acquisition, and social literacy are significantly correlated with the conversion of farmers’ e-commerce entrepreneurial behaviors. However, the practical effectiveness of advanced digital literacy remains underutilized. Policy design should therefore distinguish between two tiers: foundational popularization, which involves expanding on-site village guidance, skills workshops, and distance education to strengthen practical operations and platform usage, and advanced empowerment, which entails cautiously promoting advanced digital technology training based on regional needs to prevent mismatches between resource allocation and actual outcomes.

Second, it is crucial to establish a resource integration and sharing platform centered on online social capital. This study proves that online social capital exerts a significant mediating effect between digital literacy and entrepreneurial behavior transformation. Governments, cooperatives, and enterprises should collaborate to develop digital platforms that integrate market intelligence, technical support, and production-marketing linkages. Additionally, they should actively utilize social media to foster collaborative networks among farmers, enabling experience-sharing and resource complementarity.

In addition, it is essential to strengthen organizational support and optimize the precise allocation of policy resources. This study demonstrates that both instrumental support and institutional support play significant moderating roles. We recommend strengthening logistics and digital infrastructure, promoting full coverage of village-level parcel delivery services, and emphasizing cooperative models such as “cooperative-farmer partnerships” and “enterprise-farmer collaborations.” These models should provide technical tools, risk-sharing mechanisms, and incubation services to reduce entrepreneurial risks and enhance the success rate of behavioral transformation.

Finally, attention should be directed toward group and regional heterogeneity, with the implementation of categorized guidance and precision interventions. This study proves that in regions characterized by strong cooperative entrepreneurial preferences and well-developed rural logistics infrastructure, farmers are more likely to facilitate entrepreneurial behavior transformation through enhanced digital literacy. Therefore, digital empowerment policies should be prioritized in regions with robust cooperative foundations and advanced logistics conditions. Conversely, in areas with weaker resource endowments and inadequate logistics, priority should be given to addressing infrastructure and logistical deficiencies while fostering a collaborative atmosphere and organized operational models.

Research Limitations and Future Directions

This study has certain limitations. The cross-sectional nature of the data limits the ability to accurately capture the long-term influence of digital literacy on the intention-behavior gap in farmers’ e-commerce entrepreneurship. Furthermore, key variables like digital literacy are based on self-reported measures, potentially introducing measurement errors and social desirability bias. The analysis is also constrained by its reliance on survey data from specific Chinese regions. Future studies would benefit from using longitudinal data, adopting more objective measurement tools and diverse data sources, and expanding the research to cross-regional or cross-cultural comparisons.

Footnotes

ORCID iD

Yanling Wang

Ethical Considerations

This study was conducted in accordance with ethical standards. Informed consent was obtained from all participants, and the research utilized anonymized data that involved no identifiable private information or commercial interests.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Social Science Foundation of China (No. 21BJY227; 23ZDA055) and the Social Science Planning Project of Shandong Province (No. 23CJJJ25).

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

Data will be made available on request.

References

Chen

Hou

Liao

Wang

(2024). Digital environment, digital literacy, and farmers’ entrepreneurial behavior: A discussion on bridging the digital divide. Sustainability, 16(23), 10220.

Cofré-Bravo

Klerkx

Engler

(2019). Combinations of bonding, bridging, and linking social capital for farm innovation: How farmers configure different support networks. Journal of Rural Studies, 69, 53–64.

Cui

Pan

S. L.

Newell

Cui

(2017). Strategy, resource orchestration and e-commerce enabled social innovation in rural China. Journal of Strategic Information Systems, 26(1), 3–21.

Eisenberger

Fasolo

Davis-LaMastro

(1990). Perceived organizational support and employee diligence, commitment, and innovation. Journal of Applied Psychology, 75(1), 51–59.

Fahmi

N. A.

Zulkifli

Irwanto

Fathurohman

Pratama

I. W. A.

(2023). Analysis of the influence of e-commerce use and digital literacy toward society intention in digital entrepreneurship. MALCOM: Indonesian Journal of Machine Learning and Computer Science, 3(2), 161–167.

Fornell

Larcker

D. F.

(1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50.

George

Merrill

R. K.

Schillebeeckx

S. J. D.

(2021). Digital sustainability and entrepreneurship: How digital innovations are helping tackle climate change and sustainable development. Entrepreneurship Theory and Practice, 45(5), 999–1027.

Gilster

(1997). A primer on digital literacy. 2009. http://www.ibiblio.org/cisco/noc/primer.html

Gollwitzer

P. M.

(1999). Implementation intentions: Strong effects of simple plans. American Psychologist, 54(7), 493–503.

10.

Hair

J. F.

Black

W. C.

Babin

B. J.

Anderson

R. E.

(Eds.). (2009). Multivariate data analysis (7th ed.). Prentice-Hall.

11.

Haji

(2021). E-commerce development in rural and remote areas of BRICS countries. Journal of Integrative Agriculture, 20(4), 979–997.

12.

Hartmann

Backmann

Newman

Brykman

K. M.

Pidduck

R. J.

(2022). Psychological resilience of entrepreneurs: A review and agenda for future research. Journal of Small Business Management, 60(5), 1041–1079.

13.

Havrilova

L. H.

Topolnik

Y. V.

(2017). Digital culture, digital literacy, digital competence as the modern educational phenomena. Information Technologies and Learning Tools, 61(5), 1–14.

14.

Hayes

A. F.

(2017). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford publications.

15.

Keikhay Farzane

Sardar Shahraki

Tash

Aramesh

. (2024). Innovation solutions and business opportunity in the context of e-commerce in rural cooperatives. Innovation EconomicEcosystem Studies, 4(2), 1–17.

16.

Law

N. W. Y.

Woo

D. J.

De la Torre

Wong

K. W. G.

(2018). A global framework of reference on digital literacy skills for indicator 4.4. 2. UNESCO.

17.

Leong

Pan

Ractham

Kaewkitipong

(2015). ICT-enabled community empowerment in crisis response: Social media in Thailand flooding 2011. Journal of the Association for Information Systems, 16(3), 174–212.

18.

Liang

Casteel

Janssen

Wang

Rohlman

D. S.

(2023). Organizational resources and social support influences on stress and depression: A comparison among cooperative and non-cooperative farmers. Journal of Agromedicine, 28(2), 177–186.

19.

Zhang

Zhao

Chansanam

Song

(2025). Developing a digital literacy framework for rural farmers in China. Journal of the Australian Library and Information Association, 74(1), 93–114.

20.

Zang

Chandio

A. A.

Yang

Jiang

(2023). Farmers’ adoption of digital technology and agricultural entrepreneurial willingness: Evidence from China. Technology and Society, 73, 102253.

21.

Liu

Min

Liu

(2021). The adoption and impact of E-commerce in rural China: Application of an endogenous switching regression model. Journal of Rural Studies, 83, 106–116.

22.

Lyon

(2000). Trust, networks and norms: The creation of social capital in agricultural economies in Ghana. World Development, 28(4), 663–681.

23.

Mei

Mao

Chu

(2020). Effects and mechanisms of rural E-commerce clusters on households' entrepreneurship behavior in China. Growth and Change, 51(4), 1588–1610.

24.

Mukul

Sheeri

(2024). A study on social capital and incubation in enriching rural entrepreneurship in rural areas. Journal of Entrepreneurship & Management, 13(1), 8.

25.

Park

K. G.

Han

Kaid

L. L.

(2013). Does social networking service usage mediate the association between smartphone usage and social capital? New Media & Society, 15(7), 1077–1093.

26.

Podsakoff

P. M.

MacKenzie

S. B.

Podsakoff

N. P.

Lee

J. Y.

(2003). The mismeasure of man(agement) and its implications for leadership research. Leadership Quarterly, 14(6), 615–656.

27.

Prasandha

Susanti

Y. D.

(2022). Empowering rural entrepreneurs through independent-entrepreneurship literacy program. ASEAN Journal of Community Engagement, 6(1), 48–75.

28.

Zheng

Guo

(2019). The formation of Taobao villages in China. China Economic Review, 53, 106–127.

29.

Rhoades

Eisenberger

(2002). Perceived organizational support: A review of the literature. Journal of Applied Psychology, 87(4), 698–714.

30.

Rubyutsa

J. M.

Achtenhagen

Stendahl

Musekura

(2024). Advancing rural entrepreneurship in Rwanda through informal training – Insights from Paulo Freire’s pedagogy of the oppressed. Entrepreneurship Education and Pedagogy, 7(3), 347–376.

31.

Song

Yang

(2024). A study on the impact mechanism of internet embedding on rural E-commerce entrepreneurship. Research in International Business and Finance, 68, 102196.

32.

Tang

Zhu

(2020). Informality and rural industry: Rethinking the impacts of E-Commerce on rural development in China. Journal of Rural Studies, 75, 20–29.

33.

The World Bank. (2024). Digital progress and trends report 2023. The World Bank. https://doi.org/10.1596/978-1-4648-2049-6

34.

Wang

Tian

Nasrullah

Zhang

(2025). Does social capital influence farmers’ e-commerce entrepreneurship? China’s regional evidence. Electronic Commerce Research, 25, 3397–3425.

35.

Williams

(2006). On and off the 'Net: Scales for social capital in an online era. Journal of Computer-Mediated Communication, 11(2), 593–628.

36.

Xie

Huang

Bin

Apostolidis

Jiang

Cai

(2022). Sustainable entrepreneurship in rural E-commerce: Identifying entrepreneurs in practitioners by using deep neural networks approach. Frontiers in Environmental Science, 10, 840479.

37.

Yaşlak

Akgün

A. A.

Baycan

(2023). Social networks of online rural entrepreneurs: The case of Turkey. Annals of Regional Science, 70(3), 705–721.

38.

Zeng

Wang

Zhang

(2024). Analysis of collaborative development of e-commerce in agricultural product logistics. Journal of Theory and Practice of Management Science, 4(01), 6–10.

39.

Zhao

Lin

Knerr-Sievers

Mardani

(2023). The impact of institutional environment on entrepreneurial performance in micro E-commerce for Women: The mediating role of entrepreneurial network. Journal of Business Research, 154, 113313.

40.

Zimmerman

M. A.

(2000). Empowerment theory: Psychological, organizational and community levels of analysis. In Rappaport

Seidman

(Eds.), Handbook of community psychology (pp. 43–63). Springer.

Could Digital Literacy Transform Farmers With E-Commerce Entrepreneurial Intentions Into Actual Entrepreneurs? The Moderating Effect of Organizational Support

Abstract

Plain Language Summary

Keywords

Introduction

Intention-Behavior Gap: The Dilemma Under Multiple Constraints

Digital Literacy: Empowerment Potential

The Present Study: Delving Deeper

Literature Review and Hypotheses Development

Literature Review

Hypotheses Development

The Influence of Digital Literacy on the Intention-Behavior Gap in Farmers’ E-Commerce Entrepreneurship

The Influence of Digital Literacy on the Online Social Capital

The Influence of Online Social Capital on the Intention-Behavior Gap in Farmers’ E-commerce Entrepreneurship

The Moderating Role of Organizational Support

Methodology

Research Design and Data Collection

Measures

Dependent Variable

Independent Variables

Mediating Variables

Moderate Variables

Control Variables

Data Estimation Techniques

Baseline Regression Model

Bootstrap Approach

Data Analysis and Results

Common Method Bias

Reliability and Validity Tests

Hypothesis Testing of the Main Effects

Baseline Regression

Endogeneity Test

Mediating Effect Test

Moderating Effect Test

Robustness Test

Heterogeneity Analysis

Discussion

Conclusions and Implications

Conclusions

Implications

Implications of Theory

Implications of Practice

Research Limitations and Future Directions

Footnotes

ORCID iD

Ethical Considerations

Funding

Declaration of Conflicting Interests

Data Availability Statement

References