How Does Social Mobilization Affect Farmers’ Green Grain Production in China?

Abstract

Farmers’ adoption of green grain production practices is essential for advancing China’s ecological civilization and achieving carbon neutrality. However, adoption remains uneven because farmers’ decisions are embedded in local social structures and shaped by short-term economic incentives and constraints. Drawing on an embeddedness framework, this study investigates how social mobilization influences farmers’ green grain production practices, while also examining the moderating role of household resource endowments and the mediating role of non-market value perceptions. Using multi-stage survey data collected in Heilongjiang Province between June and September 2023, the results show that grassroots cadres foster farmers’ green production adoption through four dimensions of social mobilization—technical, knowledge, cultural, and relational embeddedness. Moreover, household endowments positively moderate these effects, and non-market value perceptions partially mediate the relationship between social mobilization and green production practices. These findings are robust to alternative model specifications. This study provides micro-level evidence on how a cadre-led, governance-based social mobilization process is associated with farmers’ adoption of green production practices. Overall, this study advances understanding of the behavioral foundations of farmers’ green transitions and highlights actionable policy levers for grassroots governance, helping translate external policy directives into internalized and sustainable production practices.

1. Introduction

Against the backdrop of China’s vigorous pursuit of ecological civilization and carbon neutrality, farmers’ adoption of green grain production practices is crucial for enhancing ecological quality and achieving high-quality development. Farmers are an essential pillar in safeguarding national food security and play a pivotal role in promoting local socioeconomic development and ecological conservation [1]. In grain production, farmers serve as the primary agents in the production process [2]. As the primary actors in grassroots agricultural production, farmers ultimately determine whether green production practices are to be adopted. Promoting their large-scale uptake is therefore a key challenge for advancing high-quality and sustainable development.

Embeddedness theory provides an important analytical lens for understanding farmers’ green grain production practices. In new economic sociology, embeddedness suggests that economic behavior is shaped by social relations rather than being purely atomized. Green production often entails higher upfront costs and potential yield reductions, which can erode short-term returns. Unsurprisingly, much of the literature on farmers’ adoption decisions centers on expected economic payoffs. Farmers are more willing to adopt behavioral measures that promise higher expected returns [3]. Yet in the current economic context, the main benefits of transitioning to green production are reflected less in immediate economic gains than in improvements in environmental quality and the fulfillment of social responsibility. From the perspective of embeddedness, farmers’ green grain production practices are shaped not only by market economic logic but also by the social structures and cultural environments in which they are situated.

Social mobilization is widely recognized as a key mechanism for policy implementation and resource coordination, and it constitutes an important pathway for effective environmental governance. At the political level, grassroots cadres transmit and implement governmental policies, serving as a direct bridge between farmers and national policy mandates. Through social mobilization, they translate national priorities into locally intelligible guidance and facilitate implementation on the ground [4]. However, when divergent interests and behavioral logics arise between farmers and governance actors, grassroots governance may become dysfunctional or less effective [5]. To prevent governance failures at the local level, governing bodies must embed governance principles into farmers’ behavioral logic through robust mobilization mechanisms, thereby enabling these principles to be internalized and translated into routine production practices. In doing so, social mobilization helps to mitigate disconnection and tension between governance actors and production agents.

Prior research suggests that farmers’ adoption of green technologies is influenced by social capital, extension services, and cognitive factors. However, these strands of scholarship typically focus either on social capital rooted in horizontally formed community networks and generalized trust, or on agricultural extension centered on the provision of technical information and training. In this study, we employ the notion of cadre-led social mobilization to examine a policy implementation and state–society linkage process. Specifically, village cadres translate policy objectives into locally actionable guidance and, through coordinated informational dissemination, norm building, and relationship maintenance, embed these objectives into farmers’ day-to-day production practices. Although this governance-centered process may incorporate extension-like elements, it cannot be reduced to extension services or generic social capital because it is actor-specific with grassroots cadres as initiators, strongly policy-oriented in emphasizing implementation and compliance, and explicitly aimed at fostering the internalization of public and ecological values.

This study advances embeddedness theory by integrating social mobilization and key actors in green production into a unified analytical framework, thereby addressing gaps in the existing literature and extending the scope of inquiry. Farmers’ adoption of green grain production practices cannot be fully accounted for by the classical utility-maximization approach. To date, few studies have systematically examined—through an embeddedness lens—how social mobilization shapes farmers’ adoption of green production practices. Drawing on micro-level survey data from 696 farming households, we estimate the impact of social mobilization on farmers’ green production practices and unpack the mechanisms implied by embeddedness. We further test the moderating effect of household endowments and the mediating role of non-market value perceptions. Overall, the study deepens understanding of farmers’ production decision-making processes and offers both theoretical and policy-relevant insights for designing effective social mobilization strategies and strengthening grassroots governance to promote green production.

2. Theoretical Analysis and Research Hypotheses

2.1. Effects of Social Mobilization on Farmers’ Green Production Practices

Within the framework of sustainable development theory, farmers’ adoption of green production practices and the extent of such adoption are not only critical for advancing the green transformation of agricultural production and management but also fundamental for maintaining long term ecosystem stability and promoting the sustainable development of the socioeconomic system. A study by Li identified several issues in farmers’ adoption of green production practices, including low motivation, limited adoption levels, and improper use of green technologies [6]. These issues pose significant obstacles to achieving sustainable development. According to rational peasant theory, farmers, as rational economic agents, make decisions about adopting green production practices based on multiple interest considerations and a tradeoff between costs and expected returns [7]. From a sustainable development perspective, this trade-off involves not only economic interests but also ecological and social considerations. Accordingly, adopting green production practices extends beyond farmers’ private economic benefits, bearing directly on ecological balance, biodiversity conservation, and the cultivation of social responsibility.

In contemporary grassroots governance, social mobilization has become a key factor shaping farmers’ green grain production practices. embeddedness is not used as a generic label for support or social ties; it specifically refers to the process by which governance objectives and policy rationales become incorporated into farmers’ day-to-day decision logics and production routines through repeated interactions in local networks. From an embeddedness perspective, farmers, as socially embedded individuals, are influenced not only by internal personal characteristics but also by the social and cultural environments in which they operate [8]. The influence of these environments is particularly pronounced under dispersed agricultural operations, where farmers often display bounded rationality and their production decisions may not fully align with broader societal goals. In this context, the role of grassroots government institutions is crucial. Grassroots authorities need to integrate farmers into local environmental governance networks through four forms of embeddedness, namely technical embeddedness, knowledge embeddedness, cultural embeddedness, and relational embeddedness. These mechanisms not only facilitate the sharing of knowledge and information but also strengthen shared community values and behavioral norms, thereby playing an important role in encouraging farmers to adopt green production practices [9]. By perceiving and participating in such embedded social mobilization, farmers gain essential knowledge and experience, which enables them to develop behavioral patterns that are consistent with broader societal objectives [10,11]. This process of social mobilization reflects both internal capacity building and external embedding. Within this dual process, grassroots governments can leverage their advantages in social mobilization to activate and guide farmers’ self-governance practices [12].

Grassroots governments, through multidimensional social mobilization strategies embedded in local social structures, have effectively promoted farmers’ adoption of green production practices. The core of these strategies lies in the integrated application of four key dimensions, namely technical embeddedness, knowledge embeddedness, cultural embeddedness, and relational embeddedness. First, technical embeddedness, as an important approach to enhancing production efficiency, plays a critical role in advancing technological innovation and improving farmers’ technical capabilities [13]. It not only strengthens connections between farmers and external markets but also generates social value for grassroots governments by supporting the refinement of technological policies and the transformation of production models [7]. Second, knowledge embeddedness within social mobilization functions as a vital mechanism for deepening farmers’ cognitive understanding of green production. The effectiveness of mobilization largely depends on farmers’ clear comprehension of governance logic at the grassroots level [14]. Through knowledge embeddedness, information related to green production practices becomes deeply rooted in farmers’ cognition, helping them to recognize both the feasibility and the long-term benefits of adopting such practices. In this process, grassroots governance relies not only on rational knowledge but also on experiential knowledge that extends beyond conventional scientific frameworks [15]. Third, cultural embeddedness emphasizes that when economic strategies and objectives are formulated, rational economic actors are shaped by shared collective understandings within their surrounding environment. The adoption of green grain production practices, with farmers as the primary actors, reflects the integration of local values, customs, and behavioral norms. Rooted in local communities, grassroots governments can better preserve and promote local cultural traditions through cultural embeddedness, thereby fostering stronger bonds with farmers. Finally, relational embeddedness refers to the establishment of mutual understanding between grassroots governments and farmers on the basis of equality and inclusiveness. This form of embeddedness reflects norms of moral respect and helps to reduce the psychological distance between the two parties [16]. The consensus built through these interactions contributes to mitigating information asymmetries in agricultural production. The grassroots autonomy shaped through relational embeddedness reflects the internalization of green production practices and facilitates farmers’ active adoption of such practices. Accordingly, the following hypotheses are proposed to illustrate the relationship between social mobilization and farmers’ green production practices.

H1. 

Social mobilization positively influences farmers’ green production practices.

H1a. 

The technical embeddedness of social mobilization has a positive effect on farmers’ green production practices.

H1b. 

The knowledge embeddedness of social mobilization has a positive effect on farmers’ green production practices.

H1c. 

The cultural embeddedness of social mobilization has a positive effect on farmers’ green production practices.

H1d. 

The relational embeddedness of social mobilization has a positive effect on farmers’ green production practices.

2.2. The Moderating Effect of Household Endowments

Household endowments play a crucial role in farmers’ production decision making. In this study, “household endowments” is defined as a multidimensional bundle of resources and capabilities that includes both economic and non-economic dimensions. In the current context of household based agricultural operations, households should be treated as the fundamental unit of production when analyzing farmers’ adoption of green production practices. For farmers, household endowments are not only a source of production inputs but also a key factor shaping their production decisions and methods. Households with higher levels of resource endowments are more likely to respond positively to social mobilization initiatives led by grassroots governments. First, the scale and composition of household endowments directly affect farmers’ ability to perceive and participate in social mobilization [17,18]. Households with greater resources and stronger capabilities are better positioned to meet the production requirements and environmental standards promoted through social mobilization. Second, household endowments also influence farmers’ response strategies when they confront market fluctuations and policy changes. The diversity and flexibility of these endowments can significantly enhance farmers’ capacity to comply with environmental protection policies and accelerate their adoption of green production practices.

Household endowments play a crucial moderating role in how social mobilization influences farmers’ adoption of green grain production practices. The level of these endowments largely determines farmers’ capacity to adopt green production practices and thus has a significant impact on their production decision making [19]. Within the framework of social practice theory, differences in individual capital endowments lead to variation in behavioral choices. Different forms of capital exhibit a certain degree of substitutability, and through the integration of multiple types of capital, dominant forms can compensate for weaker ones, thereby alleviating the constraints imposed by limited endowments. In this theoretical context, the heterogeneity of household endowments substantially shapes farmers’ responses to green grain production practices. Distinct levels of household resources give rise to differentiated capabilities and responses when farmers adjust to new production requirements and environmental standards. The mechanisms underlying this moderating effect can be summarized as follows. First, abundant household endowments significantly enhance farmers’ flexibility and adaptability in responding to social mobilization. Qiao argues that the more resources a household possesses, the more proactive and innovative it becomes when transitioning toward environmentally friendly production practices [20]. A high level of resource availability not only provides households with greater economic and social capital but also strengthens their resilience to environmental change. As a result, farmers with sufficient resource endowments are better able to respond to and internalize social mobilization initiatives led by grassroots governments. This enhanced capacity enables farmers to align more closely with governmental policy orientations when implementing sustainable development practices and thereby promotes the adoption of green grain production practices. Second, diverse household endowments not only improve farmers’ ability to adopt new technologies and practices but also deepen their understanding of green production concepts and requirements. Households with abundant resources are generally more capable of comprehending new technologies and methods related to environmental protection because they possess the economic resources for investment as well as the knowledge base and technical capabilities required to implement green production practices effectively. As household resources and capabilities increase, farmers can respond more rapidly to adjustments in production practices. Therefore, household endowments function as a key moderating factor in the relationship between social mobilization and farmers’ adoption of green production practices. Based on this, the following hypotheses are proposed:

H2. 

Household endowments have a positive moderating effect on the relationship between social mobilization and farmers’ green production practices.

H2a. 

Household endowments have a positive moderating effect on the relationship between the technical embeddedness of social mobilization and farmers’ green production practices.

H2b. 

Household endowments have a positive moderating effect on the relationship between the knowledge embeddedness of social mobilization and farmers’ green production practices.

H2c. 

Household endowments have a positive moderating effect on the relationship between the cultural embeddedness of social mobilization and farmers’ green production practices.

H2d. 

Household endowments have a positive moderating effect on the relationship between the relational embeddedness of social mobilization and farmers’ green production practices.

2.3. The Mediating Effect of Non-Market Value Perceptions

In economic research, the perception of value derived from goods and services that can be directly consumed or used through market transactions is commonly referred to as market value perception [21]. This concept is typically associated with attributes of products or services that can be bought and sold in the market, and it emphasizes their immediate economic benefits and market demand. In contrast, the perception of indirect use values that cannot be consumed or exchanged through market mechanisms is referred to as non-market value perception. Shi defines non-market value as an objectively existing form of value that cannot be realized through market transactions [22]. It generally includes evaluations of ecosystem services, biodiversity conservation, and cultural heritage preservation, as well as a deeper understanding of their roles in sustainable development and ecological civilization. Such recognition involves not only the protection and improvement of the ecological environment but also an emphasis on farmers self development and the sustainable well being of future generations [23]. There is a clear distinction between non-market value and market value, since the former cannot be expressed through conventional market exchanges. Within the framework of social cognition theory, the social environment plays a crucial role in shaping individual cognitive processes. Individual cognition is influenced not only by personal experience but also by social interactions, cultural contexts, and societal norms [24]. Therefore farmers understanding and evaluation of value are not static but evolve dynamically in response to social mobilization. When social mobilization promotes the dissemination and recognition of green production values, farmers value perceptions adjust accordingly. This transformation reflects the influence of the social environment on individual cognition and value formation. In different social contexts, farmers may re evaluate the non-market value of their products or services and may attach greater importance to social responsibility, sustainable development, and other non-economic factors in their production practices.

Social mobilization plays an important role in strengthening farmers’ perceptions of non-market values. In many cases, farmers tend to prioritize the market value of their production activities while overlooking non-market values that are equally important for society and the environment [25]. Social mobilization functions not only as a vehicle for grassroots governments to convey governance principles but also as an effective mechanism for embedding these principles in local social life. It helps to improve the alignment between the governance logic of grassroots authorities and the everyday logic of rural communities, thereby promoting mutual coordination between the two. Through social mobilization, grassroots governments can more effectively communicate to farmers the long term benefits of green production practices. As farmers internalize these messages, they come to recognize that green production practices are not only beneficial for their immediate economic interests but also have far reaching impacts on ecosystem stability, sustainable development, and the overall quality of community life.

Social mobilization can strengthen farmers’ perceptions of non-market values and thereby shape their green production practices. From the perspective of traditional economics, farmers are generally regarded as rational economic agents whose primary goal in economic activities is to pursue profit maximization and efficiency improvement [26]. As a result, their production practices are mainly driven by expected economic returns [27]. Farmers’ production decisions are influenced by both their cognitive capacities and the environments in which they operate [28]. As farmers develop a deeper understanding of the non-market values associated with green production practices, their pursuit of these values may begin to outweigh their pursuit of traditional market values [29]. This shift is reflected in a stronger tendency to adopt green production practices in actual decision making. With the advancement of social mobilization, heightened non-market value perceptions lead farmers to consider not only economic returns but also the long term environmental and social consequences of their production activities, which in turn increases their willingness to adopt green production practices. Based on this, the following hypotheses are proposed:

H3. 

Non-market value perceptions mediate the relationship between social mobilization and farmers’ green production practices.

H3a. 

Non-market value perceptions mediate the relationship between the technical embeddedness of social mobilization and farmers’ green production practices.

H3b. 

Non-market value perceptions mediate the relationship between the knowledge embeddedness of social mobilization and farmers’ green production practices.

H3c. 

Non-market value perceptions mediate the relationship between the cultural embeddedness of social mobilization and farmers’ green production practices.

H3d. 

Non-market value perceptions mediate the relationship between the relational embeddedness of social mobilization and farmers’ green production practices.

3. Data and Empirical Methods

3.1. Data Source

The data used in this study were collected through field surveys conducted in Heilongjiang Province, China, from June to September 2023. To ensure representativeness and methodological rigor, a combination of simple random sampling and stratified sampling was adopted. Specifically, in each of six sample cities, one to three counties or districts were randomly selected. Within each selected county or district, one to three townships were then chosen, and within each township one or two natural villages were included. From each village, five to ten farmers were randomly selected as respondents. This multistage sampling strategy ensured both the breadth and diversity of the sample and thereby enhanced the generalizability and reliability of the findings. In total, 850 questionnaires were distributed. After careful screening and data cleaning, questionnaires with incomplete or internally inconsistent information were removed, yielding 696 valid responses and an effective response rate of 81.88%. This high proportion of valid questionnaires supports the accuracy and credibility of the empirical results. It should be noted that our sample is drawn primarily from Heilongjiang Province, a major grain-producing region in China, which may exhibit distinctive regional features in terms of natural endowments, the intensity of policy implementation, and farming and operational structures. However, Heilongjiang is broadly representative with respect to the promotion of green grain production policies and grassroots governance practices, and the mechanisms examined in this study reflect general micro-level behavioral processes. Therefore, this regional focus is more likely to limit the external generalizability of the findings rather than to alter the empirical conclusions regarding the pathways and underlying logic through which social mobilization operates.

3.2. Variable Setting

3.2.1. Dependent Variable: Green Production Practices

This variable is measured as the number of green production practices adopted by each farmer. Specifically, six practices are considered, namely use of low toxicity pesticides, reduction in chemical fertilizer application, straw incorporation, application of pesticides according to prescribed dosages, use of organic fertilizers, and recycling of discarded pesticide containers such as bottles or bags. For each practice, a value of 1 is assigned if the respondent adopts the practice and 0 otherwise, and the dependent variable is calculated as the sum of these six indicators.

3.2.2. Independent Variable: Social Mobilization

Social mobilization is conceptualized as a multidimensional, perception-based construct reflecting how village cadres mobilize farmers through technology diffusion, knowledge transfer, norm building, and relationship maintenance. Accordingly, we operationalize social mobilization through four dimensions: technical embeddedness, knowledge embeddedness, cultural embeddedness, and relational embeddedness. In the quantitative analysis, social mobilization serves as the key explanatory variable. To capture farmers’ perceived intensity and salience of these mobilization efforts, each dimension is measured using multiple items on a five-point Likert scale. A Likert-type format is appropriate because the four dimensions represent latent perceptions rather than directly observable behaviors, and it allows respondents to translate their experiences into comparable ordinal evaluations. We adopt a five-point scale because it strikes a balance between measurement sensitivity and respondent burden in face-to-face rural surveys: it is simple enough to reduce cognitive load and response fatigue while still providing sufficient variation for empirical identification and reliability testing.

Following prior studies, we measure technical embeddedness using three reflective items that capture the extent to which relevant actors provide timely and sufficient information disclosure, deliver practical and problem-oriented technical support, and engage in continuous experience sharing that facilitates the diffusion of production know-how and the resolution of on-farm technical constraints [30,31,32,33,34]. Knowledge embeddedness is assessed with five items designed to reflect the breadth and depth of knowledge integration, focusing on participation in business and skills training, the intensity and effectiveness of policy publicity and interpretation, and the frequency/quality of knowledge transmission through both formal and informal channels [35,36,37,38]. Cultural embeddedness is measured with five items that represent the alignment of actors within a shared cultural context, including the perceived congruence of shared values, adherence to behavioral norms, and the internalization of moral responsibility, which together shape the legitimacy and acceptance of green production practices [39,40,41]. Finally, relational embeddedness is captured by four items emphasizing the strength of interpersonal and inter-organizational ties, including communication frequency, the level of mutual trust, and the propensity for help-seeking and reciprocal assistance, all of which reduce transaction frictions, enhance coordination, and support sustained engagement in collective actions related to green production [42,43,44,45]. All measurement items are summarized in

Table 1. Social mobilization indicator system.

Dimension	Measurement Items	Assignment
Technical Embeddedness	How do you perceive the extent to which the village cadres disclose information related to green agricultural production technologies? (T1)	Very little = 1, Relatively little = 2, Moderate = 3, Relatively much = 4, Very much = 5
	How do you perceive the level of support and assistance provided by the village cadres regarding green production technologies? (T2)
	How do you perceive the extent of experience sharing and exchange by the village cadres regarding green agricultural production technologies? (T3)
Knowledge Embeddedness	How do you perceive the frequency with which the village cadres provide training on green agricultural production knowledge? (K1)	Very little = 1, Relatively little = 2, Moderate = 3, Relatively much = 4, Very much = 5
	How do you perceive the intensity of the village cadres’ publicity and promotion of environmental protection policies? (K2)
	How do you perceive the intensity of the village cadres’ efforts in promoting green agricultural production? (K3)
	How often do you participate in village meetings related to ecological and environmental protection? (K4)
	Are you familiar with the national environmental protection regulations related to agricultural production? (K5)
Cultural Embeddedness	Do you share the same vision and values for agricultural development as those promoted by the village cadres? (C1)	Strongly disagree = 1, Disagree = 2, Neutral = 3, Agree = 4, Strongly agree = 5
	Do you follow the same agricultural production codes of conduct advocated by the village cadres? (C2)
	Do environmentally damaging behaviors in agricultural production make you feel uneasy? (C3)
	Do you feel guilty about environmental damage caused by agricultural production? (C4)
	Do you believe you should ensure environmental protection in agricultural production? (C5)
Relational Embeddedness	Do you maintain close contact with the village cadres? (R1)	Very little = 1, Relatively little = 2, Moderate = 3, Relatively much = 4, Very much = 5
	How frequently do you participate in the village’s Party-building activities? (R2)
	What is your level of trust in the village cadres? (R3)
	When you encounter problems, are you willing to seek help from the village cadres? (R4)

, which reports the specific indicators used for each construct.

Because social mobilization is composed of multiple dimensions and the corresponding indicators are measured by multiple Likert items, we aggregate them into continuous indices for the empirical analysis. Specifically, we first standardize the item responses and then use the entropy method to construct the composite social mobilization index and the four dimensional indices. Compared with an equal-weighting approach, entropy weighting reduces subjectivity and assigns larger weights to indicators with greater information content, improving comparability when dimensions contain different numbers of items. The weights of various dimensions of social mobilization are presented in

Table A1. Weights of social mobilization indicators.

Dimension	Measurement Items	Weights
Technical Embeddedness (0.3755)	T1	0.3229
	T2	0.3319
	T3	0.3452
Knowledge Embeddedness (0.2463)	K1	0.2362
	K2	0.1818
	K3	0.2771
	K4	0.1417
	K5	0.1632
Cultural Embeddedness (0.1227)	C1	0.2623
	C2	0.2372
	C3	0.1875
	C4	0.1726
	C5	0.1405
Relational Embeddedness (0.2554)	R1	0.1516
	R2	0.1613
	R3	0.5271
	R4	0.1601

in the Appendix A. The resulting indices are normalized to facilitate interpretation and are reported as an index scaled between 0 and 1.

3.2.3. Moderating Variable: Household Endowment

The measurement of household endowments follows Yuan [19] and uses the entropy method to construct a composite index. In this study, household endowments are conceptualized as a multidimensional household resource base that captures natural resources, human capital, financial capacity, and market-oriented production. Detailed indicator definitions and coding are provided in

Table 2. Description of household endowment variables.

Variable	Variable Definition	Assignment
Household Endowment	Planting Scale	Less than 5 mu = 1, 5–10 mu = 2, 10–20 mu = 3, 20 mu or more = 4
	Years of Farming Experience	Less than 2 years = 1, 2–5 years = 2, 5–10 years = 3, 10 years or more = 4
	Specialization in Agricultural Production (Measured as the ratio of agricultural production income to total household income.)	≦20% = 1, 20–40% = 2, 40–60% = 3, 60–80% = 4, >80% = 5
	Total Household Income	10,000 CNY
	Sales Channels	Collective unified sales = 1, Purchase by merchants = 2, Sales through intermediaries = 3

. Planting scale, years of farming experience, total household income, specialization in agricultural production, and sales channels proxy natural resources, human capital, financial capacity, market production orientation, respectively.

3.2.4. Mediating Variable: Non-Market Value Perception

This variable captures the intrinsic factors that influence farmers’ participation in green production practices. Non-market value perceptions reflect the evaluative channel through which farmers appraise the outcomes of green practices. Specifically, they capture farmers’ beliefs about the ecological benefits and intergenerational public benefits generated by adopting green practices, such as reducing pollution, improving soil quality, and enhancing long-term well-being. In essence, non-market value perceptions are value-oriented and focus on farmers’ assessments of public and ecological welfare that cannot be fully realized through market transactions. Following Shi [22], it comprises three components, namely perception of ecological value, perception of the social security value of cultivated land for the present generation, and perception of the social security value of cultivated land for future generations. Perception of ecological service value is measured by the degree of agreement with the statement that green production practices can reduce environmental pollution. Perception of social security value for the present generation is measured by the degree of agreement with the statement that green production practices can improve the quality of cultivated land. Perception of social security value for future generations is measured by the degree of agreement with the statement that green production practices are beneficial to the long-term livelihood of future generations. All responses are recorded on a five-point Likert scale. To quantify non-market value perceptions, exploratory factor analysis was conducted in SPSS 19.0. Exploratory factor analysis (EFA) was applied to reduce dimensionality after the reliability and validity of the scale data had been tested. The results showed that the Kaiser Meyer Olkin (KMO) value was 0.729, exceeding the threshold of 0.5, and that Bartlett’s test of sphericity yielded an approximate chi square value of 1614.329 (p < 0.001), indicating statistical significance. These results suggest that the data exhibit good structural validity and are suitable for factor analysis. Using principal component analysis (PCA) to determine the number of factors and applying varimax rotation with eigenvalues of at least 1 as the criterion, one common factor was extracted, which accounted for 85.868% of the total variance.

3.2.5. Control Variables

In this study, control variables are selected along three dimensions: individual characteristics, household characteristics, and regional environmental characteristics. The individual level control variables capture the potential influence of personal factors on the outcome variables. This category includes six variables: gender, age, smartphone use, health status, educational attainment, and pension insurance coverage. These variables are chosen based on their established importance in the literature on individual behavior and decision making. Gender, age, and smartphone use may affect individuals’ risk preferences and decision-making patterns, whereas educational attainment and health status reflect their capacity to access resources and their overall quality of life. Pension insurance coverage captures the level of social security protection. For household characteristics, the analysis considers the influence of the household environment on individual behavior. Three variables are included in this dimension: household support burden, total household size, and housing area. These variables are intended to reveal how family responsibilities and resource allocation shape individuals’ choices and priorities. The household support burden may influence work and life decisions by either expanding or constraining specific social and economic opportunities. With respect to regional environmental characteristics, distance to the nearest market is used as a proxy for the broader macroeconomic and social environment in which individuals reside. This distance may affect individuals’ access to and use of production and market related resources. Detailed definitions of all variables and their descriptive statistics are presented in

Table 3. Variable definitions and descriptive statistics analysis.

Variable	Definition	Mean	Std.	Minimum	Maximum
Green Production Practices	Number of Adopted Green Grain Production Practices	3.060	1.977	0.000	6.000
Social Mobilization	Calculated Using the Entropy Method	0.711	0.228	0.051	1.000
Technical Embeddedness	Calculated Using the Entropy Method	0.705	0.277	0.000	1.000
Knowledge Embeddedness	Calculated Using the Entropy Method	0.710	0.242	0.000	1.000
Cultural Embeddedness	Calculated Using the Entropy Method	0.790	0.192	0.000	1.000
Relational Embeddedness	Calculated Using the Entropy Method	0.686	0.243	0.000	1.000
Non-Market Value Perception	Obtained Through Factor Analysis	0.000	1.000	−4.334	0.949
Household Endowment	Calculated Using the Entropy Method	0.412	0.199	0.000	0.922
Gender	Male = 1, Female = 2	1.339	0.474	1.000	2.000
Age	Age	46.777	10.862	21.000	69.000
Smartphone Usage	Whether the respondent uses a smartphone: Yes = 1, No = 0	0.974	0.159	0.000	1.000
Health Status	Very poor = 1, Poor = 2, Fair = 3, Good = 4, Very good = 5	3.726	0.995	1.000	5.000
Years of Education	Number of years of formal schooling	11.633	3.892	0.000	19.000
Dependency Burden	Proportion of household members under 15 and over 65 years old to the total household population	0.355	0.352	0.000	1.000
Housing Area	Actual residential area of the household	72.945	29.106	22.000	225.000
Pension Insurance	Whether the respondent has pension insurance: Yes = 1, No = 0	0.569	0.496	0.000	1.000
Market Distance	Time required to reach the nearest market town	28.17	22.932	1.000	120.000
Household Size	Number of members living in the same household	3.503	1.422	1.000	8.000

.

3.3. Empirical Model

3.3.1. OLS Regression Model

This study integrates the theoretical analysis, research hypotheses, and micro level data on farmers’ green production practices into a coherent analytical framework. Our dependent variable measures the intensity of green practice adoption and is constructed as the sum of six binary practice indicators, yielding a bounded count from 0 to 6. Because this outcome is an additive adoption-intensity index, we first estimate a linear conditional-mean model using ordinary least squares (OLS), which directly quantifies how social mobilization changes the expected number of practices adopted and yields coefficients that are straightforward to interpret. To address the concern that the bounded and discrete nature of the outcome may induce heteroskedasticity, all OLS specifications are estimated with heteroskedasticity robust standard errors. To further examine how household endowments shape the effect of social mobilization on farmers’ green production practices, an interaction term between household endowments and social mobilization is incorporated into the model. This specification allows an assessment of the moderating role of household endowments in the relationship between social mobilization and farmers’ green production practices, thereby providing a more nuanced understanding of this relationship. Specifically, the baseline empirical model is specified as follows.

$${\mathrm{Y}}_{\mathrm{i}}={\mathsf{\alpha }}_0{+\mathsf{\alpha }}_1{\mathrm{X}}_{\mathrm{i}}+{\mathsf{\alpha }}_2{\mathrm{P}}_{\mathrm{i}}+{\mathsf{\alpha }}_3{\mathrm{X}}_{\mathrm{i}}\times {\mathrm{P}}_{\mathrm{i}}+{\mathsf{\alpha }}_4{\mathrm{E}}_{\mathrm{i}}+{\mathsf{\epsilon }}_{\mathrm{i}}$$

(1)

In Equation (1), the level of farmers’ green production practices is treated as a continuous variable. In the model, ${\mathrm{Y}}_{\mathrm{i}}$ denotes the level of green production practices of the ith farmer and serves as the dependent variable. ${\mathrm{X}}_{\mathrm{i}}$ denotes social mobilization, the key independent variable, and ${\mathsf{\alpha }}_1$ is its estimated coefficient. ${\mathrm{P}}_{\mathrm{i}}$ represents household endowments, with ${\mathsf{\alpha }}_2$ as the corresponding estimated coefficient. The interaction term ${\mathrm{X}}_{\mathrm{i}}\times {\mathrm{P}}_{\mathrm{i}}$ captures the joint effect of social mobilization and household endowments, and ${\mathsf{\alpha }}_3$ is its estimated coefficient, which is used to test how this interaction influences farmers’ adoption of green production practices. ${\mathrm{E}}_{\mathrm{i}}$ denotes the vector of control variables, and ${\mathsf{\alpha }}_4$ is the associated coefficient vector. ${\mathsf{\alpha }}_0$ is the constant term, and ${\mathsf{\epsilon }}_{\mathrm{i}}$ is the random disturbance term that captures other unobserved factors not explicitly included in the model.

3.3.2. Mediation Effect Model

When examining the impact of social mobilization on farmers’ green production practices, this study posits that the effect may operate through a mediating variable, namely farmers’ non-market value perceptions. In this framework, non-market value perceptions are treated as a key psychological construct that acts as a bridge between social mobilization and farmers’ green production practices. Specifically, farmers’ non-market value perceptions may be shaped by social mobilization and, in turn, influence the extent to which they adopt green production practices. To systematically test the mediating role of non-market value perceptions in the relationship between social mobilization and farmers’ green production practices, this study follows the mediation testing procedure proposed by Tian [46] and specifies the following empirical models.

$$Y_i=c{X}_i+{\epsilon }_{1i}$$

(2)

$$M=a{X}_i+{\epsilon }_{2i}$$

(3)

$$Y_i=c{X}_i+b{M}_i+{\epsilon }_{3i}$$

(4)

In this framework, the mediating effect is defined as the product of the coefficients $\mathrm{a}\times \mathrm{b}$, where $\mathrm{a}$ denotes the effect of the independent variable ${\mathrm{X}}_{\mathrm{i}}$ on the mediating variable $\mathrm{M}$, and $\mathrm{b}$ denotes the effect of the mediating variable $\mathrm{M}$ on the dependent variable ${\mathrm{Y}}_{\mathrm{i}}$. The terms ${\mathsf{\epsilon }}_{1\mathrm{i}}$, ${\mathsf{\epsilon }}_{2\mathrm{i}}$, and ${\mathsf{\epsilon }}_{3\mathrm{i}}$ denote the residuals of the regression equations, capturing the unexplained variance in each model, and the coefficient $\mathrm{c}$ represents the direct effect of the independent variable ${\mathrm{X}}_{\mathrm{i}}$ on the dependent variable ${\mathrm{Y}}_{\mathrm{i}}$. The analysis first examines the direct effect of the independent variable ${\mathrm{X}}_{\mathrm{i}}$ on the dependent variable ${\mathrm{Y}}_{\mathrm{i}}$, and then estimates the effect of ${\mathrm{X}}_{\mathrm{i}}$ on the mediating variable $\mathrm{M}$ as well as the effect of $\mathrm{M}$ on ${\mathrm{Y}}_{\mathrm{i}}$. This analytical procedure enables a more precise identification and quantification of the indirect pathway through which social mobilization influences farmers’ green production practices via the mediating mechanism of non-market value perceptions.

3.3.3. Binary Probit Model

In the robustness analysis, farmers’ adoption of green production practices is treated as a binary outcome variable. The variable takes the value 1 if a farmer adopts green production practices and 0 otherwise. Given the binary nature of this dependent variable, a probit model is employed to conduct the robustness analysis and to examine the effect of social mobilization on farmers’ adoption of green production practices. The probit model is well suited to binary outcomes and provides insights into the determinants of farmers’ decisions regarding the adoption of green production practices. By estimating the impact of the explanatory variables on the probability of adoption, the probit model offers an effective framework for understanding and interpreting farmers’ green production practices. The probit specification is given as follows.

$$P(Y_i=1|X_i)=\Phi ({\alpha }_0+{\alpha }_1{X}_i+{\alpha }_2{C}_i+{\epsilon }_i)$$

(5)

In this model, $\mathsf{\Phi }\left(\mathrm{*}\right)$ denotes the standard normal cumulative distribution function. ${\mathrm{Y}}_{\mathrm{i}}$ denotes the green production practices of the ith farmer and serves as the binary dependent variable. A value of 1 indicates that the farmer has adopted green production practices, whereas a value of 0 indicates non adoption. ${\mathrm{X}}_{\mathrm{i}}$ represents the degree of social mobilization experienced by the ith farmer and serves as the main independent variable. It is constructed from four embeddedness dimensions: technical embeddedness, knowledge embeddedness, cultural embeddedness, and relational embeddedness. ${\mathrm{C}}_{\mathrm{i}}$ denotes the vector of control variables that may influence the green production practices of the ith farmer. ${\mathsf{\epsilon }}_{\mathrm{i}}$ is the random disturbance term, which captures unobserved factors that are not explicitly included in the model. ${\mathsf{\alpha }}_0$, ${\mathsf{\alpha }}_1$, and ${\mathsf{\alpha }}_2$ are parameters to be estimated. This modeling approach provides an empirical basis for evaluating the impact of social mobilization on farmers’ adoption of green production practices.

4. Analysis of Empirical Results

All data analyses are conducted in Stata 15.0. The dependent variable—the level of green production practices—is treated as a continuous variable, with higher values indicating a greater degree of engagement in green production practices. Prior to the regression analysis, all continuous variables were winsorized at the 1 percent level to mitigate the influence of outliers. Moderating effects are tested using interaction terms. Before estimating the regression models, pairwise correlation analysis was conducted for the key variables, and the results are reported in

Table 4. Matrix of correlation coefficients between variables.

	Green Production Practices	Social Mobilization	Non-Market Value Perception	Household Endowment	Gender	Age	Health Status
Green Production Practices	1
Social Mobilization	0.223 ***	1
Non-Market Value Perception	0.212 ***	0.538 ***	1
Household Endowment	0.150 ***	−0.084 **	0.068 *	1
Gender	−0.042	−0.074 *	−0.063 *	−0.147 ***	1
Age	0.076 **	−0.054	−0.063 *	0.077 **	0.022	1
Health Status	0.071 *	0.306 ***	0.222 ***	0.013	−0.098 ***	−0.034	1
Years of Education	−0.009	0.110 ***	0.081 **	−0.204 ***	0.050	−0.096 **	0.259 ***
Smartphone Usage	0.138 ***	0.128 ***	0.160 ***	0.037	−0.017	0.019	0.037
Dependency Burden	0.00200	0.010	−0.041	−0.041	−0.021	−0.041	0.024
Housing Area	−0.0150	−0.002	0.036	0.068 *	−0.048	−0.020	0.221 ***
Pension Insurance	0.0250	0.098 ***	0.039	−0.186 ***	0.017	0.020	−0.053
Market Distance	0.065 *	−0.064	−0.001	0.162 ***	−0.016	0.011	−0.129 ***
Household Size	0.033	−0.020	0.007	0.221 ***	0.018	−0.042	0.061
	Years of Education	Smartphone Usage	Dependency Burden	Housing Area	Pension Insurance	Market Distance	Household Size
Years of Education	1
Smartphone Usage	0.117 ***	1
Dependency Burden	−0.070 *	−0.006	1
Housing Area	0.133 ***	−0.001	−0.029	1
Pension Insurance	0.086 **	0.114 ***	−0.018	−0.138 ***	1
Market Distance	−0.113 ***	−0.003	−0.010	−0.084 **	0.005	1
Household Size	0.052	0.009	0.006	0.270 ***	−0.156 ***	0.087 **	1

Note: ***, **, and * represent the significance levels of 1%, 5%, and 10%, respectively.

. To prevent potential multicollinearity, all variables were mean centered before constructing the interaction terms. In addition, variance inflation factor (VIF) diagnostics were performed, and the maximum VIF among all variables was 1.20, well below the conventional threshold of 10, indicating that multicollinearity is not a concern.

4.1. Baseline Regression

To further examine how social mobilization influences farmers’ green production practices, an ordinary least squares (OLS) regression model is employed. The model examines not only the overall embeddedness of social mobilization but also the effects of its four specific dimensions—technical embeddedness, knowledge embeddedness, cultural embeddedness, and relational embeddedness—on farmers’ green production practices. By incorporating these dimensions as independent variables, the analysis provides a comprehensive perspective on how different aspects of social mobilization shape farmers’ adoption of green production practices. The regression results are reported in

Table 5. Regression results of the effect of social mobilization on the adoption of green production practices.

	(1)	(2)	(3)	(4)	(5)
Social Mobilization	1.746 ***
	(4.88)
Technical Embeddedness		1.359 ***
		(4.67)
Knowledge Embeddedness			1.494 ***
			(4.49)
Cultural Embeddedness				2.133 ***
				(5.15)
Relational Embeddedness					1.190 ***
					(3.54)
Gender	−0.002	−0.012	−0.022	−0.018	−0.005
	(−0.02)	(−0.07)	(−0.14)	(−0.11)	(−0.03)
Age	0.131 *	0.126 *	0.128 *	0.130 *	0.133 *
	(1.88)	(1.80)	(1.83)	(1.87)	(1.89)
Health Status	0.076	0.094	0.085	0.096	0.122
	(0.90)	(1.12)	(1.00)	(1.16)	(1.45)
Years of Education	0.016 *	0.015	0.016 *	0.016 *	0.017 *
	(1.66)	(1.57)	(1.66)	(1.71)	(1.79)
Smartphone Usage	1.202 **	1.252 **	1.312 **	1.159 **	1.243 **
	(2.25)	(2.34)	(2.46)	(2.17)	(2.30)
Dependency Burden	−0.033	−0.028	−0.016	−0.009	−0.032
	(−0.20)	(−0.17)	(−0.10)	(−0.05)	(−0.19)
Housing Area	−0.005	−0.006	−0.010	−0.007	−0.018
	(−0.07)	(−0.08)	(−0.12)	(−0.09)	(−0.22)
Pension Insurance	0.055	0.082	0.082	0.088	0.050
	(0.35)	(0.52)	(0.52)	(0.56)	(0.31)
Market Distance	0.148 *	0.150 *	0.145 *	0.141 *	0.145 *
	(1.89)	(1.92)	(1.86)	(1.80)	(1.84)
Household Size	−0.008	−0.011	−0.009	−0.015	0.001
	(−0.14)	(−0.18)	(−0.15)	(−0.25)	(0.02)
N	696	696	696	696	696

Note: ***, **, and * represent the significance levels of 1%, 5%, and 10%, respectively.

. Columns (1), (2), (3), (4), and (5), respectively, present the estimated effects of social mobilization, technical embeddedness, knowledge embeddedness, cultural embeddedness, and relational embeddedness on farmers’ green production practices. The findings show that all embeddedness variables are statistically significant at the 1 percent level, indicating that social mobilization and its dimensions have a significant positive effect on farmers’ adoption of green production practices. In other words, higher levels of social mobilization are associated with a greater degree of engagement in green production practices. These empirical results highlight the important role of social mobilization in promoting farmers’ green production practices and provide empirical support for H1.

4.2. Test of Moderating Effects

The analysis further examines the moderating role of household endowments in farmers’ adoption of green production practices. As shown in

Table 6. Regression results for the moderating effect of household endowments.

	(1)	(2)	(3)	(4)	(5)
Social Mobilization	1.046 **
	(2.37)
Social Mobilization × Household Endowment	1.759 ***
	(2.80)
Technical Embeddedness		0.709 *
		(1.82)
Technical Embeddedness × Household Endowment		1.561 **
		(2.53)
Knowledge Embeddedness			0.812 *
			(1.90)
Knowledge Embeddedness × Household Endowment			1.694 ***
			(2.73)
Cultural Embeddedness				1.507 ***
				(3.08)
Cultural Embeddedness × Household Endowment				1.587 ***
				(2.86)
Relational Embeddedness					0.451
					(1.04)
Relational Embeddedness × Household Endowment					1.870 ***
					(2.85)
Gender	0.056	0.040	0.034	0.049	0.051
	(0.36)	(0.25)	(0.22)	(0.31)	(0.32)
Age	0.118 *	0.116 *	0.115	0.115 *	0.118 *
	(1.69)	(1.66)	(1.64)	(1.66)	(1.67)
Health Status	0.066	0.086	0.075	0.090	0.111
	(0.75)	(0.99)	(0.85)	(1.05)	(1.27)
Years of Education	0.010	0.010	0.011	0.011	0.012
	(1.02)	(1.00)	(1.06)	(1.08)	(1.14)
Smartphone Usage	1.189 ***	1.254 ***	1.302 ***	1.151 ***	1.210 ***
	(2.89)	(3.04)	(3.17)	(2.85)	(2.91)
Dependency Burden	−0.014	−0.010	0.004	0.011	−0.012
	(−0.08)	(−0.06)	(0.02)	(0.07)	(−0.07)
Housing Area	−0.014	−0.015	−0.018	−0.011	−0.028
	(−0.18)	(−0.19)	(−0.24)	(−0.14)	(−0.37)
Pension Insurance	0.110	0.128	0.134	0.140	0.112
	(0.70)	(0.82)	(0.86)	(0.90)	(0.69)
Market Distance	0.111	0.116	0.111	0.103	0.110
	(1.35)	(1.42)	(1.35)	(1.26)	(1.34)
Household Size	−0.028	−0.028	−0.028	−0.036	−0.019
	(−0.46)	(−0.46)	(−0.46)	(−0.59)	(−0.31)
N	696	696	696	696	696

Note: ***, **, and * represent the significance levels of 1%, 5%, and 10%, respectively.

, the interaction terms between household endowments and the various dimensions of social mobilization are all statistically significant. First, the interaction between overall social mobilization and household endowments is statistically significant at the 1 percent level, indicating a positive moderating effect of household endowments on the relationship between social mobilization and farmers’ green production practices. Second, the interaction between technical embeddedness and household endowments is statistically significant at the 5 percent level, suggesting that household endowments positively moderate the effect of technical embeddedness on farmers’ green production practices. For knowledge embeddedness, its interaction with household endowments is significant at the 1 percent level, confirming a positive moderating effect of household endowments on the relationship between knowledge embeddedness and farmers’ green production practices. Similarly, the interaction between cultural embeddedness and household endowments is statistically significant at the 1 percent level, indicating that household endowments exert a positive moderating effect in this relationship. Finally, the interaction between relational embeddedness and household endowments is statistically significant at the 1 percent level, further demonstrating that household endowments positively moderate the impact of relational embeddedness on farmers’ adoption of green production practices. Taken together, these findings provide empirical support for H2. Based on the regression coefficients from models (1)–(5), the marginal enhancement effect of household endowments on green production practices, as farmers’ perceptions of social mobilization increase, follows the order relational embeddedness > knowledge embeddedness > cultural embeddedness > technical embeddedness.

4.3. Test of Mediation Mechanism

This section reports the results of testing the mediating role of non-market value perceptions in the relationship between social mobilization and farmers’ green production practices.

Table 7. Test of the mediating effect of non-market value perceptions between social mobilization on the adoption of green production practices.

	(1)		(2)		(3)		(4)
	Non-Market Value Perception	Green Production Practices	Non-Market Value Perception	Green Production Practices	Non-Market Value Perception	Green Production Practices	Non-Market Value Perception	Green Production Practices
Non-Market Value Perception		0.289 ***		0.286 ***		0.227 **		0.333 ***
		(3.37)		(3.25)		(2.24)		(4.03)
Technical Embeddedness	1.447 ***	0.941 ***
	(7.96)	(3.03)
Knowledge Embeddedness			1.963 ***	0.933 **
			(10.06)	(2.50)
Cultural Embeddedness					3.077 ***	1.434 ***
					(13.43)	(2.65)
Relational Embeddedness							1.543 ***	0.676 *
							(8.67)	(1.88)
Gender	−0.062	0.007	−0.066	−0.003	−0.056	−0.005	−0.044	0.010
	(−0.83)	(0.04)	(−0.91)	(−0.02)	(−0.87)	(−0.03)	(−0.56)	(0.06)
Age	−0.055	0.142 **	−0.051	0.143 **	−0.046	0.141 **	−0.045	0.148 **
	(−1.63)	(2.07)	(−1.58)	(2.07)	(−1.47)	(2.05)	(−1.30)	(2.13)
Health Status	0.096 **	0.066	0.058	0.068	0.059 *	0.083	0.108 ***	0.086
	(2.49)	(0.76)	(1.55)	(0.78)	(1.81)	(0.96)	(2.75)	(0.98)
Years of Education	−0.001	0.016	−0.001	0.016 *	0.000	0.017 *	0.001	0.017 *
	(−0.29)	(1.63)	(−0.13)	(1.70)	(−0.03)	(1.73)	(0.27)	(1.77)
Smartphone Usage	0.535	1.097 ***	0.559 *	1.153 ***	0.317	1.087 ***	0.472	1.086 ***
	(1.63)	(2.65)	(1.82)	(2.77)	(1.10)	(2.68)	(1.34)	(2.59)
Dependency Burden	−0.081	−0.004	−0.075	0.006	−0.068	0.007	−0.095	0.000
	(−1.13)	(−0.03)	(−1.09)	(0.03)	(−1.13)	(0.04)	(−1.33)	(0.00)
Housing Area	0.016	−0.011	0.018	−0.015	0.024	−0.013	0.007	−0.020
	(0.40)	(−0.14)	(0.44)	(−0.19)	(0.72)	(−0.16)	(0.16)	(−0.26)
Pension Insurance	0.024	0.075	0.009	0.079	0.010	0.085	−0.031	0.060
	(0.34)	(0.48)	(0.13)	(0.51)	(0.15)	(0.55)	(−0.43)	(0.38)
Market Distance	0.048	0.136 *	0.044	0.133	0.037	0.132	0.044	0.131
	(1.40)	(1.69)	(1.29)	(1.64)	(1.32)	(1.64)	(1.25)	(1.61)
Household Size	−0.029	−0.003	−0.028	−0.001	−0.038	−0.007	−0.014	0.006
	(−1.06)	(−0.05)	(−1.07)	(−0.02)	(−1.62)	(−0.11)	(−0.54)	(0.10)
N	696	696	696	696	696	696	696	696

Note: ***, **, and * represent the significance levels of 1%, 5%, and 10%, respectively.

presents the empirical results of the mediation analysis. Detailed estimates of the direct effects of social mobilization and its embedded dimensions on farmers’ green production practices have already been reported in the baseline regression and are therefore not repeated here. Column (1) shows that technical embeddedness has a significant positive effect on farmers’ non-market value perceptions, which is statistically significant at the 1 percent level. When both technical embeddedness and non-market value perceptions are included in the model, non-market value perceptions continue to exert a significant positive effect on farmers’ green production practices at the 1 percent level, and the effect of technical embeddedness remains positive and statistically significant at the 1 percent level. This pattern indicates that non-market value perceptions play a partial mediating role between technical embeddedness and farmers’ green production practices. Column (2) indicates that knowledge embeddedness has a significant positive effect on non-market value perceptions and is statistically significant at the 1 percent level. When both knowledge embeddedness and non-market value perceptions are entered into the model, non-market value perceptions maintain a significant positive influence on farmers’ green production practices at the 1 percent level, while the effect of knowledge embeddedness remains positive and statistically significant at the 5 percent level, suggesting a partial mediation effect of non-market value perceptions between knowledge embeddedness and farmers’ green production practices. Column (3) demonstrates that cultural embeddedness has a significant positive effect on farmers’ non-market value perceptions, with significance at the 1 percent level. When both cultural embeddedness and non-market value perceptions are included in the model, non-market value perceptions continue to have a significant positive effect on farmers’ green production practices at the 1 percent level, and the effect of cultural embeddedness remains positive and statistically significant at the 1 percent level, confirming a partial mediation effect. Column (4) shows that relational embeddedness also has a significant positive effect on non-market value perceptions and is statistically significant at the 1 percent level. When both relational embeddedness and non-market value perceptions are included in the model, non-market value perceptions retain a significant positive effect on farmers’ green production practices at the 1 percent level, and the effect of relational embeddedness remains positive and statistically significant at the 10 percent level, indicating a partial mediation effect between relational embeddedness and farmers’ green production practices. In summary, non-market value perceptions exhibit a mediating effect through which social mobilization promotes farmers’ adoption of green production practices. These results therefore provide empirical support for H3.

4.4. Robustness Test

To assess the robustness and reliability of the conclusions drawn from the baseline analysis, this section reports a series of robustness checks, including additional efforts to address potential reverse causality. First, we re-estimate the baseline specification by transforming the dependent variable from the number of adopted green production practices into a binary indicator of whether a farmer adopts green production practices. We then estimate a binary Probit model to test whether the effect of social mobilization remains stable under an alternative model specification, which helps mitigate concerns that the main results are driven by functional form choices. In addition, we re-examine the moderating role of household endowments under this alternative setting. The corresponding results are presented in

Table 8. Robustness test of social mobilization on the adoption of green production practices by farmers.

	(1)	(2)	(3)	(4)	(5)
Social Mobilization	1.913 ***
	(6.18)
Technical Embeddedness		1.476 ***
		(5.79)
Knowledge Embeddedness			1.620 ***
			(5.59)
Cultural Embeddedness				1.538 ***
				(4.20)
Relational Embeddedness					1.557 ***
					(5.47)
Gender	−0.045	−0.039	−0.065	−0.082	−0.047
	(−0.34)	(−0.30)	(−0.50)	(−0.64)	(−0.36)
Age	0.055	0.051	0.053	0.054	0.054
	(0.99)	(0.92)	(0.96)	(0.97)	(0.98)
Health Status	−0.013	0.001	−0.005	0.040	0.019
	(−0.19)	(0.02)	(−0.07)	(0.59)	(0.28)
Years of Education	0.007	0.006	0.008	0.009	0.009
	(0.97)	(0.84)	(1.00)	(1.21)	(1.24)
Smartphone Usage	−0.297 **	−0.236	−0.252 *	−0.222	−0.369 **
	(−2.01)	(−1.62)	(−1.73)	(−1.52)	(−2.41)
Dependency Burden	−0.062	−0.059	−0.048	−0.039	−0.070
	(−0.47)	(−0.44)	(−0.37)	(−0.30)	(−0.53)
Housing Area	−0.020	−0.020	−0.030	−0.023	−0.024
	(−0.31)	(−0.31)	(−0.46)	(−0.35)	(−0.38)
Pension Insurance	0.156	0.152	0.165	0.184	0.172
	(1.16)	(1.14)	(1.24)	(1.40)	(1.27)
Market Distance	0.091	0.091	0.090	0.087	0.091
	(1.53)	(1.54)	(1.52)	(1.49)	(1.54)
Household Size	0.085 *	0.085 *	0.083 *	0.079 *	0.091 **
	(1.84)	(1.82)	(1.79)	(1.71)	(1.99)
N	696	696	696	696	696

Note: ***, **, and * represent the significance levels of 1%, 5%, and 10%, respectively.

and

Table 9. Robustness tests for the moderating effect of household endowments.

	(1)	(2)	(3)	(4)	(5)
Social Mobilization	0.554 *
	(1.87)
Social Mobilization × Household Endowment	1.217 ***
	(3.05)
Technical Embeddedness		0.311
		(1.20)
Technical Embeddedness × Household Endowment		1.092 ***
		(2.79)
Knowledge Embeddedness			0.314
			(1.12)
Knowledge Embeddedness × Household Endowment			1.096 ***
			(2.76)
Cultural Embeddedness				1.013 ***
				(3.07)
Cultural Embeddedness × Household Endowment				1.142 ***
				(3.14)
Relational Embeddedness					0.339
					(1.19)
Relational Embeddedness × Household Endowment					1.328 ***
					(3.24)
Gender	−0.095	−0.103	−0.108	−0.094	−0.095
	(−0.84)	(−0.92)	(−0.97)	(−0.82)	(−0.85)
Age	−0.004	−0.004	−0.005	−0.005	−0.004
	(−0.09)	(−0.08)	(−0.10)	(−0.10)	(−0.09)
Health Status	0.032	0.045	0.046	0.038	0.050
	(0.55)	(0.79)	(0.80)	(0.66)	(0.88)
Years of Education	0.000	0.000	0.000	0.000	0.000
	(−0.06)	(−0.04)	(0.05)	(0.02)	(0.04)
Smartphone Usage	−0.168	−0.137	−0.138	−0.129	−0.214 *
	(−1.35)	(−1.11)	(−1.12)	(−1.03)	(−1.68)
Dependency Burden	0.000	0.003	0.012	0.012	−0.003
	(−0.00)	(0.03)	(0.11)	(0.11)	(−0.03)
Housing Area	−0.016	−0.017	−0.020	−0.012	−0.022
	(−0.29)	(−0.31)	(−0.37)	(−0.21)	(−0.40)
Pension Insurance	0.293 ***	0.296 ***	0.299 ***	0.297 ***	0.296 ***
	(2.59)	(2.62)	(2.65)	(2.62)	(2.61)
Market Distance	0.067	0.070	0.067	0.064	0.065
	(1.23)	(1.29)	(1.24)	(1.17)	(1.21)
Household Size	−0.016	−0.015	−0.014	−0.024	−0.012
	(−0.39)	(−0.36)	(−0.33)	(−0.57)	(−0.29)
N	696	696	696	696	696

Note: ***, **, and * represent the significance levels of 1%, 5%, and 10%, respectively.

. Second, to test the robustness of the mediating mechanism, we conduct hierarchical regression analyses to evaluate both the existence and magnitude of the mediating effect of non-market value perceptions in the link between social mobilization and farmers’ green production practices. This procedure further clarifies the channel through which social mobilization shapes farmers’ adoption behavior and the mediating contribution of non-market value perceptions. The detailed results are reported in

Table 10. Robustness tests for the mediating effect of non-market value perceptions.

	Green Production Practices	Non-Market Value Perception	Green Production Practices
Non-Market Value Perception			0.267 ***
			(3.02)
Social Mobilization	1.746 ***	2.120 ***	1.180 ***
	(4.82)	(10.13)	(2.93)
Gender	−0.002	−0.045	0.010
	(−0.02)	(−0.63)	(0.06)
Age	0.131 *	−0.048	0.144 **
	(1.88)	(−1.47)	(2.09)
Health Status	0.076	0.058	0.060
	(0.87)	(1.57)	(0.69)
Years of Education	0.016	−0.001	0.016 *
	(1.64)	(−0.14)	(1.69)
Smartphone Usage	1.202 ***	0.441	1.084 ***
	(2.97)	(1.38)	(2.62)
Dependency Burden	−0.033	−0.094	−0.008
	(−0.20)	(−1.37)	(−0.05)
Housing Area	−0.005	0.021	−0.011
	(−0.07)	(0.53)	(−0.14)
Pension Insurance	0.055	−0.017	0.060
	(0.35)	(−0.25)	(0.38)
Market Distance	0.148 *	0.046	0.135 *
	(1.81)	(1.42)	(1.67)
Household Size	−0.008	−0.027	−0.001
	(−0.14)	(−1.03)	(−0.02)
N	696	696	696

Note: ***, **, and * represent the significance levels of 1%, 5%, and 10%, respectively.

. Overall, the robustness checks yield consistent coefficient signs and statistical significance patterns across specifications, suggesting that the findings are not sensitive to alternative measurements or model forms and that potential reverse causality is unlikely to materially bias the main conclusions.

The probit model is first used to estimate the effect of social mobilization on whether farmers adopt green production practices, and the corresponding regression results are reported in detail. These results are summarized in Table 8, where columns (1), (2), (3), (4), and (5), respectively, present the effects of social mobilization, technical embeddedness, knowledge embeddedness, cultural embeddedness, and relational embeddedness on farmers’ green production practices. As shown in Table 8, all embeddedness dimensions are statistically significant at the 1 percent level, indicating that each has a significant positive effect on farmers’ green production practices. This finding suggests that social mobilization and its embedded dimensions play an important role in promoting farmers’ adoption of green production practices and that social mobilization helps motivate farmers to transition toward more sustainable production methods. The probit-based robustness results are consistent with both the theoretical framework and the earlier empirical findings, thereby reinforcing the overall conclusions and providing additional empirical support for H1.

Furthermore, a robustness analysis is conducted to examine the moderating role of household endowments in the relationship between social mobilization and farmers’ adoption of green production practices. The analysis focuses on how household endowments moderate the relationship between each dimension of social mobilization and farmers’ green production practices. As shown in Table 9, the interaction terms between household endowments and each dimension of social mobilization are all statistically significant. First, the interaction between overall social mobilization and household endowments is statistically significant at the 1 percent level, indicating that household endowments exert a positive moderating effect on the relationship between the degree of social mobilization and farmers’ green production practices. Second, the interaction between technical embeddedness and household endowments is also statistically significant at the 1 percent level. For knowledge embeddedness, its interaction with household endowments is statistically significant at the 1 percent level, and the interaction between cultural embeddedness and household endowments is likewise statistically significant at the 1 percent level, indicating that household endowments play a positive moderating role in the relationship between cultural embeddedness and farmers’ green production practices. Finally, the interaction between relational embeddedness and household endowments is statistically significant at the 1 percent level, underscoring the positive moderating effect of household endowments in this relationship. Overall, these findings show that the robustness results are consistent with the previous empirical analyses and thereby reinforce the reliability of the study’s conclusions. The results provide additional empirical support for H2.

This analysis further investigates the mediating effect of non-market value perceptions in the relationship between social mobilization and farmers’ adoption of green production practices, and the corresponding robustness results are reported in Table 10. The regression results show that, in Model (1), social mobilization has a significant positive effect on farmers’ green production practices. In Model (2), social mobilization also has a significant positive effect on farmers’ non-market value perceptions. In Model (3), when both social mobilization and non-market value perceptions are included in the regression, the effect of social mobilization on farmers’ green production practices remains positive and statistically significant at the 1 percent level, even after controlling for non-market value perceptions. This finding indicates that non-market value perceptions partially mediate the relationship between social mobilization and farmers’ green production practices. The consistency between these results and the earlier empirical findings reinforces the robustness of the study’s conclusions and provides additional empirical support for H3.

5. Conclusions and Policy Recommendations

5.1. Research Conclusions

Drawing on survey data, this study employs OLS regression and mediation models to empirically examine the impact of social mobilization on farmers’ green production practices. The analysis considers not only the direct effects of social mobilization on farmers’ green production practices but also the moderating role of household endowments and the mediating mechanism of non-market value perceptions. The main findings are summarized as follows.

First, social mobilization and its four embedded dimensions are positively and significantly associated with farmers’ adoption of green production practices. Specifically, technical, knowledge, cultural, and relational embeddedness each increase the likelihood of adoption. Robustness checks using a Probit specification yield consistent results, confirming the stability of the baseline findings.

Second, household endowments positively moderate the relationship between social mobilization and farmers’ green production practices. This indicates that farmers with higher household endowments are more likely to adopt green production practices under the influence of social mobilization. The robustness analysis based on the probit model further supports this moderating effect. The coefficients for social mobilization and its embedded dimensions, as well as for the interaction terms with household endowments, remain statistically significant, providing additional empirical support for the hypothesis that social mobilization positively influences farmers’ adoption of green production practices.

Third, non-market value perceptions significantly mediate the relationship between social mobilization and farmers’ green production practices. This mediation holds not only for overall social mobilization but also for each of its four embedded dimensions—technical, knowledge, cultural, and relational embeddedness. Specifically, non-market value perceptions partially mediate the effects of these dimensions, indicating that social mobilization promotes farmers’ engagement in green production by strengthening their recognition of non-market values. Mediation robustness checks yield qualitatively unchanged and statistically significant coefficients, supporting the stability of the findings.

5.2. Policy Recommendations

Based on the research conclusions, the following policy recommendations are proposed.

1.

Strengthening multi-dimensional social mobilization to promote green production.

Given that social mobilization and its four embedded dimensions, namely technical embeddedness, knowledge embeddedness, cultural embeddedness, and relational embeddedness, are all significantly and positively associated with farmers’ adoption of green production practices, policy efforts should prioritize the development of a multi level and multi channel mobilization framework that facilitates a transition from externally driven compliance to internally motivated engagement. First, policies should reinforce technical embeddedness by expanding agricultural technology demonstration sites and improving extension and service systems, thereby enhancing farmers’ capacity to access, understand, and apply green production technologies. Second, knowledge embeddedness can be strengthened through targeted technical training, field workshops, and farmer field school programs, which help deepen farmers’ understanding of green production concepts, practical techniques, and their long term benefits. Third, cultural embeddedness should be promoted by integrating ecological civilization principles into rural governance and community development, fostering shared norms and value orientations that support the diffusion and persistence of green production practices. Finally, relational embeddedness should be leveraged by activating local social networks such as village cadres, farmer cooperatives, and demonstration households to establish trust based channels for information exchange, knowledge sharing, and collective action.

2.

Enhancing household endowments to improve the effectiveness of social mobilization.

Given that household endowments play a significant and positive moderating role in the relationship between social mobilization and farmers’ adoption of green production practices, policy efforts should focus on reducing resource disparities among farmers and improving participation conditions for households with lower endowments, thereby enhancing the overall effectiveness of social mobilization. First, the fiscal support system should be strengthened by expanding green credit instruments, policy-based agricultural insurance, and targeted fiscal subsidies to provide financial security and risk protection for resource constrained farmers. Second, investments in vocational education and agricultural skills training should be increased to enhance human capital, thereby improving farmers’ knowledge base, technical capacity, and farm management capabilities. Third, asset integration should be promoted through institutional arrangements such as land transfer, cooperative farming, and agricultural machinery sharing, which can improve production efficiency and strengthen farmers’ capacity to withstand economic and environmental risks. By improving household resource endowments, social mobilization can generate broader, more equitable, and more durable incentive effects across different groups of farmers.

3.

Fostering non-market value perceptions to sustain green production practices.

Social mobilization influences farmers’ behavior not only through external incentives or institutional constraints, but also by strengthening the recognition and internalization of ecological and public values, thereby promoting both the adoption and persistence of green production practices. Accordingly, policy design should place greater emphasis on value cultivation and environmental awareness building, encouraging a gradual shift in farmers’ behavior from externally enforced compliance toward internally motivated commitment. First, ecological value education should be strengthened by systematically integrating core concepts such as environmental protection, the public good nature of ecological resources, and intergenerational responsibility into agricultural extension services and technical training programs. Second, community participation and collaborative governance should be promoted through institutional arrangements including ecological action initiatives, village regulations, and participatory environmental monitoring, in order to reinforce farmers’ sense of collective ecological responsibility. Finally, non-market incentive mechanisms should be established and improved, including green certification schemes, eco labeling programs, and public recognition initiatives, which confer social legitimacy and reputational rewards on environmentally responsible production practices. By continuously cultivating non-market value perceptions, policymakers can stimulate more stable and enduring intrinsic motivation, thereby enhancing the long term consistency and sustainability of green production behavior.