How Livelihood Capital Affects Farmers’ Green Production Behavior: Analysis of Mediating Effects Based on Farmers’ Cognition

Abstract

In light of global climate change and sustainable agricultural growth, it is critical to look at producers’ green production methods. Enhancing the quality of agricultural goods and reducing agricultural pollution are the main goals of future agricultural growth, and this is accomplished by farmers using green production methods. Regarding the research data of 364 vegetable farmers, this study uses structural equation modeling and a mediation effect model to empirically assess the effect of livelihood capital and farmers’ cognition on their green behavior. The results show that (1) natural capital, human capital, financial capital, and social capital in the livelihood capital of vegetable growers may significantly impact producers’ green production behavior. It is not immediately clear how physical capital affects the way green manufacturing practices work. (2) Natural capital and green production behavior are completely mediated by environmental and policy cognition, while human capital and green production behavior are partly mediated by environmental and policy cognition. Human and natural capital indirectly influence farmers’ sustainable production techniques via these activities. Financial and social capital directly influence farmers’ sustainable production methods, with no mediation effect seen. (3) Farmers’ green production behavior is more significantly influenced by their cognitive behavior than by their livelihood capital. Accordingly, it is recommended that environmental education and policy promotion be strengthened, that farmers’ livelihood capital be accumulated via a variety of channels, that farmers’ subsidies for green production be increased, and that farmers’ knowledge of green production be improved. The cognitive level of farmers should also be raised. In addition to providing theoretical justification for analyzing farmers’ green production practices within the framework of sustainable agricultural development, this study also acts as a guide for pertinent government agencies to help farmers choose more ecologically friendly farming methods.

1. Introduction

Green agricultural development is regarded as the remedy for agricultural environmental pollution, safeguarding the quality and safety of agricultural products while protecting the ecological environment in response to the significant challenges presented by global climate change and the pressing necessity for sustainable agricultural advancement [1]. For an extended period, the agricultural economy’s rapid expansion was facilitated by the crude growth model, which was dependent on the continuous input of resource factors. However, this model also resulted in the degradation of agricultural products’ quality and the accumulation of chemical residues, which posed a significant threat to human health [2]. Additionally, it caused issues such as soil nutrient loss, crop yield reduction, and severe agricultural surface pollution [3], which directly impacted and constrained the sustainable development of agriculture. According to the second national pollution source census bulletin, the percentages of total nitrogen (TN, 1.41 million tons), agricultural chemical oxygen demand (CODcr, 10.67 million tons), and agricultural surface source pollution (TN, 1.5 million tons) are all equal. Nitrogen (TN) emissions accounted for 1.41 million tons, while total phosphorus (TP) emissions accounted for 0.21 million tons. These emissions accounted for 49.77%, 46.4%, and 66.6% of the total emissions, respectively [4]. In 2022, the fertilizer and pesticide utilization rates of China’s three primary cereal crops—rice, wheat, and maize—will be 41.3% and 41.8%, respectively, as per the China Agricultural Green Development Report of 2023. The total quantity of fertilizer and pesticide used will decrease while efficiency will increase. However, the intensity of application is much higher than the international safety limit [5,6], and the gap is still significant when compared to 50–65% in developed nations (Department of Science, Technology and Education, Ministry of Agriculture and Rural Development, 2021). As the main body responsible for agricultural production activities, it is imperative that it look into farmers’ green production behavior and its influencing factors in order to support farmers’ participation in green production, protect the rural ecological environment, and realize the sustainable development of agriculture.

Academics have examined and debated the elements that encourage producers to practice green production from a range of angles. Numerous research studies have shown that financial capital has a major impact on producers’ green production behavior in terms of livelihood capital [7]. Producers are more likely to be able to get the funding required to support their green production techniques when the economy is doing well [8]. The ability of producers to take risks is also enhanced. The tendency of producers to invest in organic fertilizers and use green pest control methods is significantly and favorably influenced by the degree of education of human capital and the overall number of family workers. This tendency to give priority to the development of environmentally friendly industrial processes is the result of these reasons [9]. Furthermore, compared to those in traditional agriculture, green agricultural producers are more likely to be younger and female [10]. Nonetheless, some scholars still support the other position. For instance, it was discovered that older people were more likely to invest more time in green production, and that education level either had no effect on green production [11] or had a negative association with it [12]. Farmers’ green production practices may also be influenced by natural capital, irrigation and hydration conditions, and the extent of arable land fragmentation [13]. Social trust and social involvement in terms of social capital have a favorable impact on producers’ adoption of green production technology [14]. It is clear that livelihood capital has a considerable effect on green production practices [15,16,17]. Producers are more likely to implement a range of green production practices when their capital endowment is higher [18].

Ecological cognitive inequalities arise as a consequence of producers’ varying levels of green cognition, which are influenced by the availability of knowledge on green production [19]. This further limits manufacturers’ options for environmentally friendly manufacturing. This implies that producers are more likely to adopt green production practices if they have a greater degree of green cognition [20,21]. This is illustrated by the fact that the selection of pesticide application behavior by farmers is substantially influenced by their risk perception, pesticide cognition, and pest prevention and control cognition [22,23]. Empirical studies have also shown that producers can identify the economic, social, and ecological benefits of green production by improving their comprehension of the concept. As a result, they are more inclined to take the initiative to acquire the requisite technology and information, thereby decreasing their dependence on pesticides and fertilizers. Furthermore, scholars have examined the direct and indirect impacts of these factors on farmers’ propensity and behavior to participate in green agricultural production [24,25] and have integrated farmers’ cognition with social interaction and environmental regulation [26]. These studies all substantiate the beneficial influence of farmers’ cognition on the promotion of their involvement in green production.

In summary, this investigation has acquired a wealth of valuable references as a result of the more comprehensive research conducted by academicians on the green production practices of farmers, both domestically and internationally. Nevertheless, there are still certain deficiencies. First, more research has been conducted on how livelihood capital affects farmers’ green agricultural practices; nevertheless, most of these studies concentrate on one or more aspects of livelihood capital without carrying out a thorough assessment from the standpoint of livelihood capital as a whole. Second, while researchers are interested in the ways that farmers’ cognition influences their green production behavior, few have integrated livelihood capital, farmers’ cognition, and farmers’ green production behavior into a single analytical framework to investigate the relative contributions of various aspects of farmers’ cognition to the relationship between livelihood capital and green production behavior, e.g., how many facets of farmers’ thinking function as intermediaries in the relationship between their livelihood capital and green production methods.

Therefore, this paper selects vegetable farmers as the study’s focus, analyzes and assesses the direct or indirect effects of livelihood capital and farmers’ mindset on their green production practices, and obtains a scientific basis for government organizations to support, guide, and regulate the implementation of relevant laws pertaining to farmers’ green production practices.

2. Theoretical Framework and Research Hypotheses

2.1. Theoretical Frameworks

In order to develop a theoretical analysis framework for the influence of livelihood capital on farmers’ green production behavior, this study integrates the sustainable livelihood analysis framework put forth by the UK Department for International Development (DFID) with the actual circumstances in the study area (Figure 1). According to the theory of sustainable livelihoods, farmers’ choices about productivity are greatly influenced by their livelihood capital, which serves as the foundation for livelihood activities and is a crucial component in the selection of their livelihood strategies. The framework states that the basis for farmers’ participation in agricultural production activities is their livelihood capital. Farmers’ decisions to use green production methods also vary depending on the differences in livelihood capital. Furthermore, a good overall position of livelihood capital would encourage farmers to embrace green production techniques and vice versa, demonstrating that both high and low levels of livelihood capital have a direct influence on farmers’ green production practices. Additionally, the framework states that farmers’ livelihood capital affects their thoughts and actions. Farmers’ green production behavior will be indirectly impacted by changes in their livelihood capital, which will lead to changes in their cognition, as shown in Figure 1.

2.2. Research Hypotheses

2.2.1. Impact of Livelihood Capital on Farmers’ Green Production Behavior

Farmers’ livelihood capital is classified into five categories by the framework for sustainable livelihood analysis: human capital, social capital, financial capital, natural capital, and physical capital. All of the fundamental natural resources that farmers depend on to sustain their livelihoods, including the quality of cultivation space and arable land, are regarded as natural capital. In general, producers are more inclined to engage in green production, the yield is higher, and the cost of green production is lower when the cultivation area is larger, and the quality of their arable land is greater [27]. The human capital of a farm is the total labor force status, which encompasses the number of employees, their bodily condition, and their skill set. This capital is frequently determined by factors such as age or education. When there are more workers, better health, and more knowledge and skills on the farm, it is easier to master green production technology and, therefore, change green production behavior. Studies have shown that both the number of workers and their cultural background greatly increase the efficiency of agricultural production inputs; when environmental protection is a concern, workers who are more culturally aware are more likely to adopt green production practices [28]. The many instruments and materials that farmers require for daily tasks and agricultural production, among other things, make up the majority of physical capital. The adoption of reduction-oriented green production techniques is supported by farmers’ increased production efficiency, which is made possible by enough physical capital. Financial resources may be used by farmers to support both their agricultural pursuits and their families. If farmers have more money, they can buy costlier production inputs like soil testing fertilizer, organic fertilizer, biopesticides, etc. Furthermore, they are more inclined to implement environmentally friendly agricultural production methods and are more adept at addressing hazards [29]. Throughout production and management operations, producers establish a network of connections with numerous social issues, which is referred to as social capital. The links between farmers, their social standing, their engagement in the community, and other factors are the main representations of this network. Social network interactions that strengthen social identity and promote neighborly trust may motivate farmers to embrace green technologies [30]. It is clear that farmers with greater livelihood capital are more inclined to use green production methods in the agricultural production process. Based on the aforementioned data, the study’s hypothesis is as follows (as shown in

Table 1. Hypotheses related to livelihood capital on farmers’ green production behavior.

Hypotheses	Hypothetical Content
H1a:	Natural capital has a considerable positive impact on the green production behavior of farmers
H1b:	Material capital has a considerable positive impact on the green production behavior of farmers
H1c:	Human capital has a considerable positive impact on the green production behavior of farmers
H1d:	Financial capital has a considerable positive impact on the green production behavior of farmers
H1e:	Social capital has a considerable positive impact on the green production behavior of farmers

):

H1: 

The green production behavior of producers is significantly positively influenced by livelihood capital.

2.2.2. Influence of Farmers’ Perception on Farmers’ Green Production Behavior

In accordance with the theory of farmers’ cognitive behavior [31], their cognition and behavior are mutually influenced, as farmers are the primary producers of agricultural products. The concepts, beliefs, self-perception, and other cognitive aspects of agricultural families are more clearly elucidated by the two study components of environmental cognition and policy cognition. The knowledge and understanding of the agricultural production environment, which includes awareness of natural conditions such as soil, water source, and climate, as well as awareness of resource consumption, pollution, and other issues in the agricultural production process, is the main definition of environmental cognition for farm households. The level of the environmental cognition of farmers will determine their production behavior choices, including whether they choose green production practices. Farmers’ green production techniques include external issues such as social cognition and policy in addition to internal considerations. Farmers’ understanding of and compliance with rules are key factors in their adoption of green farming methods. Policy cognition is the process of comprehending and understanding agricultural policies, including their objectives, contents, methods of implementation, and outcomes. Subsequently, policy cognition substantially influences producers’ green production strategies. Given that it is evident that the cognition of farmers is significantly influenced by their livelihood capital endowment and that their cognition significantly influences their decision to participate in green production, the following hypotheses are proposed in this study:

H2: 

Environmental cognition has a considerable positive impact on farmers’ green production behavior.

H3: 

Policy cognition has a considerable positive impact on farmers’ green production behavior.

2.2.3. Mediating Impact of Farmers’ Cognition in Livelihood Capital and Farmers’ Green Production Behavior

The livelihood capital of different farmers varies, and this has an impact on their cognitive abilities. With an emphasis on the sustainable development of individuals, society, and the environment, livelihood capital serves as the tangible basis and guarantee for farmers’ survival and expansion. According to cognitive behavioral theory, human behavior is influenced by both external events and one’s internal cognition or way of thinking. As farmers’ knowledge of environmental issues and agricultural regulations increases, their cognitive level also increases, which further impacts their choice to choose green production methods. This implies that environmental and policy cognition mediates the relationship between livelihood capital and farmers’ green production practices. Accordingly, this research proposes the following hypotheses (see

Table 2. Hypotheses related to livelihood capital on farmers’ environmental cognition and policy cognition.

Hypotheses	Hypothetical Content
H4a:	Natural capital has a considerable positive impact on environmental cognition
H4b:	Material capital has a considerable positive impact on environmental cognition
H4c:	Human capital has a considerable positive impact on environmental cognition
H4d:	Financial capital has a considerable positive impact on environmental cognition
H4e:	Social capital has a considerable positive impact on environmental cognition
H5a:	Natural capital has a considerable positive impact on policy cognition
H5b:	Material capital has a considerable positive impact on policy cognition
H5c:	Human capital has a considerable positive impact on policy cognition
H5d:	Financial capital has a considerable positive impact on policy cognition
H5e:	Social capital has a considerable positive impact on policy cognition
H6a:	Environmental cognition has a mediating impact on the relationship between farmers’ natural capital and farmers’ green production behavior
H6b:	Environmental cognition has a mediating impact on the relationship between farmers’ material capital and farmers’ green production behavior
H6c:	Environmental cognition has a mediating effect on the relationship between farmers’ human capital and farmers’ green production behavior
H6d:	Environmental cognition has a mediating impact on the relationship between farmers’ financial capital and farmers’ green production behavior
H6e:	Environmental cognition has a mediating impact on the relationship between farmers’ social capital and farmers’ green production behavior
H7a:	Policy cognition mediates the impact of the relationship between farmers’ natural capital and farmers’ green production behavior
H7b:	Policy cognition mediates the impact of the relationship between farmers’ material capital and farmers’ green production behavior
H7c:	Policy cognition mediates the impact of the relationship between farmers’ human capital and farmers’ green production behavior
H7d:	Policy cognition mediates the impact of the relationship between farmers’ financial capital and farmers’ green production behavior
H7e:	Policy cognition mediates the impact of the relationship between farmers’ social capital and farmers’ green production behavior

):

H4: 

Livelihood capital has a considerable positive impact on environmental cognition.

H5: 

Livelihood capital has a considerable positive impact on policy cognition.

H6: 

Environmental cognition plays a mediating impact in the relationship between farmers’ livelihood capital and farmers’ green production behavior.

H7: 

Policy cognition mediates the impact of the relationship between farmers’ livelihood capital and farmers’ green production behavior.

3. Variable Selection and Model Construction

3.1. Study Area and Data Sources

The province of Gansu has unique natural features. The province’s various natural geographic characteristics include wide daily temperature variations, a dry environment, plenty of sunlight, extended daylight hours, and low rates of illness and insect activity. The vegetables cultivated here are of exceptional quality, slow to mature, rich in nutrients, and have a rich, flavorful taste. The main ingredients of the agricultural products referred to as “Ganwei” are these veggies. Following extensive development, six strategically advantageous vegetable-producing zones have been established in the Gannan Tibetan Area, the Hexi Corridor Irrigation Area, the Central and Along the Yellow Irrigation Area, the Jinghe River Basin, the Weihe River Basin, the “Two Rivers and One Water” Basin, and the Jinghe River Basin. As a result, it has become a crucial element in China’s efforts to encourage vegetable trade between its northern and southern provinces as well as between its eastern and western regions. It was formally acknowledged by the Ministry of Agriculture in 2008 as one of the leading vegetable-growing and exporting districts in the interior northwest of China. It is now recognized as one of China’s five main hubs for exporting northern veggies to the south and selling western crops to the east.

Since Yuzhong and Wushan are Gansu Province’s two most productive vegetable-producing districts, the committee chose them as sample locations. The stratified sampling method was used to choose 16 villages and 8 townships (

Table 3. Regional distribution of the survey.

County	Township	Village	Number of Questionnaires	Number of Questionnaires	Ratio (%)
Yuzhong county	Peace Town	Shen Jiahe Village	22
		Yuanjiaying Village	22
	Xiaguanying Town	Gaodunying Village	23
		Xiaguanying Village	22	179	49.18%
	Gancaodian Town	Sandunying Village	21
		Xicun Village	23
	Dingyuan Town	Anjiaying Village	22
		Xiejiazui Village	24
Wushan county	Chengguan Town	Qingchi Village	25
		Honggou Village	24
	Luomen Town	Wensi Village	22
		Tama Village	22	185
	Shandan Town	Zhouzhuang Village	23		50.82%
		Shandan Village	24
	Mali Town	Beishun Village	23
		Fu Men Village	22
Total	8	16	364		100

). The financial resources of vegetable growers, the use of fertilizer for vegetable growth, and the physical management and preservation of the study area were the main topics of the team’s fieldwork in January and February 2024. 364 legitimate surveys out of 400 were retrieved after erroneous answers were eliminated, resulting in an effective rate of 91%. Although the sample farmers grow a broad range of vegetables, leek, broccoli, cauliflower, cucumber, and cabbage are the most often planted. The following were the personal characteristics of the sample: The attendance rate was 47.43% female and 52.57% male. Most were in their 40s and 50s, and most had only completed elementary and junior high school. The following were the traits of the household: According to the sample, 57.6% of households had four or more persons, and 58% of families earned between CNY 20,000 and 30,000 annually per capita.

3.2. Variable Selection

3.2.1. Dependent Variable

The dependent variable chosen was Green Production Behavior, or GPB. The main ways that green production behavior is shown throughout the vegetable-growing process are via the use of fertilizers, drugs, waste treatment, and other techniques. This study, therefore, measured the indicators from four aspects: “application of organic fertilizers, green prevention and control, scientific use of medicines, and pesticide waste recycling” using the Likert 5-level scale: “very inconsistent = 1, inconsistent = 2, general = 3, more consistent = 4, and very consistent = 5). In order to assign values to the indicators, which mainly measured the four aspects of “organic fertilizer application, green control and prevention, scientific medication, pesticide waste recycling and disposal,” this study employed a 5-level Likert scale: “very non-compliant = 1, non-compliant = 2, general = 3, more compliant = 4, very compliant = 5.” On the other hand, physicochemical-induced air technology, also known as green prevention and control, has not yet gained momentum in the research region and has to be improved.

Table 4. Variable meaning and assignment.

Variable Name	Variable Definition	Average	Standard Deviation
Dependent Variable Farmers’ green production behavior (GPB) Organic fertilizer application behavior	(GPB1) Use of organic fertilizer (farmyard manure) instead of chemical fertilizer	3.19	1.205
Green prevention and control	(GPB2) Use of physical and chemical trapping and control techniques (insecticide lamps, insect nets)	3.11	1.319
Scientific use of pesticides	(GPB3) Apply pesticides according to standard dosages	3.21	1.242
Disposal of pesticide and fertilizer waste	(GPB4) Recycling and disposal of pesticide, fertilizer, and other packaging waste according to requirements	3.23	1.188
Independent Variables Natural capital (N)	(N1) Accessibility (very inconvenient = 1; less convenient = 2; average = 3; convenient = 4; very convenient = 5)	3.15	1.442
	(N2) Cultivated land area (cultivated land area per capita)	3.22	1.472
	(N3) Land output rate (output value per unit area)	3.32	1.467
Physical capital (P)	(P1) Investment in agricultural machinery and equipment (amount of agricultural machinery purchased)	4.42	1.524
	(P2) Housing status (1 room = 1; 2 rooms = 2; 3 rooms = 3; 4 rooms = 4; 5 rooms and above = 5)	2.54	1.524
Human capital (H)	(H1) Household labor force (1 and below = 1; 2–3 = 2; 3–4 = 3; 4–5 = 4; 5 and above = 5)	3.37	1.449
	(H2) Educational attainment of the main labor force (illiterate = 1; elementary = 2; middle school = 3; high school = 4; college and above = 5)	3.12	1.613
Financial capital (F)	(F1) Annual household income (annual household income)	2.88	1.041
	(F2) Household savings (less than 10,000 = 1; 10,000–30,000 = 2; 30,000–50,000 = 3; 50,000–80,000 = 4; 80,000 and above = 5)	2.84	1.052
	(F3) Difficulty of getting a loan for agricultural production (very difficult = 1; more difficult = 2; average = 3; easy = 4; very easy = 5)	2.91	1.11
Social capital (S)	(S1) Ease or difficulty of participating in skills training (very difficult = 1; more difficult = 2; average = 3; easy = 4; very easy = 5)	3.27	1.042
	(S2) Neighborhood relations (very poor = 1; poorer = 2; average = 3; better = 4; very good = 5)	3.34	0.961
	(S3) Ease of access to agricultural policy information (very difficult = 1; more difficult = 2; average = 3; easy = 4; very easy = 5)	3.26	1.015
Environmental cognition (EC)	(EC1) Fertilizer Substitution Perceptions	3.24	1.518
	(EC2) Awareness of pesticide use	3.3	1.42
	(EC3) Awareness of physical and chemical control technologies	3.33	1.487
	(EC4) Pesticide, fertilizer packaging randomly discarded hazard intensity	3.2	1.487
Policy cognition (PC)	(PC1) Awareness of rural environmental policies	3.36	1.294
	(PC2) Cognition of green production policy among farmers	3.35	1.484

makes this clear by demonstrating that all four of the green production tendencies had mean values higher than 3. Pesticide and chemical fertilizer waste treatment, scientific medication technology, green prevention and control technology, and organic fertilizer application behavior had mean values of 3.19, 3.11, 3.21, and 3.23, respectively.

3.2.2. Core Independent Variables

Livelihood Capital (LC) serves as the principal explanatory variable. The sustainable livelihood framework quantifies livelihood capital in accordance with five dimensions: natural capital, human capital, social capital, financial capital, and physical capital (hereafter denoted by N, H, S, F, and P, respectively). The accessibility, cultivated land area, and land output rate of the land are the measurements used to quantify natural capital. The education level of primary laborers and the quantity of household laborers are the determinants of human capital. Three indicators are used to quantify social capital: the ease of access to information on agricultural policies, neighborhood relations, and the ease of participation in skills training. The quantity of investment in housing conditions and agricultural apparatus and equipment is used to quantify physical capital. The annual income of the household is used to quantify financial capital. The ease of agricultural production loans and family resources. The entropy value method is employed to determine the weight of the primary independent variable of each indicator allocated to the situation, as illustrated in Table 4. Three indicators are assessed. Natural capital, physical capital, human capital, financial capital, and physical capital have mean values of 3.23, 3.48, 3.245, 2.89, and 3.29, respectively, as shown in

Table 5. Reliability analysis of each variable.

Dimension	Indicator	Corrected Term to Total Correlation	Cronbach Alpha After Deletion of Terms	Cronbach Alpha
Natural capital (N)	N1	0.64	0.754	0.808
	N2	0.678	0.714
	N3	0.651	0.743
Physical capital (P)	P1	0.673	——	0.805
	P2	0.673	——
Human capital (H)	H1	0.584	——	0.735
	H2	0.584	——
Financial capital (F)	F1	0.689	0.813	0.848
	F2	0.749	0.755
	F3	0.711	0.793
Social capital (S)	S1	0.763	0.758	0.853
	S2	0.756	0.768
	S3	0.66	0.855
Environmental cognition (EC)	EC1	0.764	0.837	0.88
	EC2	0.677	0.87
	EC3	0.77	0.834
	EC4	0.751	0.842
Policy cognition (PC)	PC1	0.578	——	0.728
	PC2	0.578	——
Green production behavior (GPB)	GPB1	0.678	0.795	0.84
	GPB 2	0.711	0.78
	GPB 3	0.686	0.792
	GPB 4	0.619	0.82

. Each variable has a mean value greater than 3, with the exception of financial capital. It is clear that the subsistence capital of farmers is in a relatively optimistic state; however, there is still room for development in terms of financial capital.

3.2.3. Mediating Variable

The mediator variable in this paper is Farmers’ Cognition (FC). Cognitive behavior theory proposes that the cognitive behavior of farmers acts as a mediator and regulator between subjective consciousness norms and behavior, emphasizing the interaction between individual subjective factors and the external environment. As previously mentioned, this paper chose to evaluate the cognition of producers using Environmental Cognition (EC) and Policy Cognition (PC). The scientific application of fertilizers and pesticides, physical and chemical-induced control technology, and the hazardousness of pesticide and fertilizer packaging are the four dimensions on which Environmental Cognition is evaluated. The environmental cognition was evaluated on four dimensions: the scientific application of fertilizers and pesticides, physicochemical control technology, the hazards of pesticides and fertilizer packaging waste, and policy cognition on two dimensions: the cognition of rural environmental policy and the cognition of farmers’ green production policy. In order to designate values to the indexes, the 5-level Likert scale was employed: “strongly disagree = 1, disagree = 2, generally agree = 3, relatively agree = 4, and strongly agree = 5”. Table 4 demonstrates that the grassroots village committees were more effective in publicizing the pertinent policies, and farmers were more conversant with the policies related to green production and had a higher sense of accord with environmental protection, with the mean values of environmental awareness and policy awareness being 3.27 and 3.36, respectively.

3.3. Model Setting

This study employs structural equation modeling to examine the factors that influence producers’ adoption of environmentally friendly production techniques. Structural equation modeling (SEM) is a versatile extension of the fundamental linear model that can be applied to a diverse array of variables. Its primary objective is to investigate the structural relationship between latent variables [32]. Structural equation modeling is divided into two categories: structural models and measurements. The structural model is typically used to investigate the relationship between latent variables and employs the following specific expression:

$$\eta =\beta \eta +\Gamma \xi +\zeta$$

(1)

In Equation (1), $\beta$ denotes the relation between endogenous latent variables; $\Gamma$ denotes the effect of exogenous latent variables on endogenous latent variables, and is the regression coefficient of $\eta$ to $\xi$; and $\zeta$ is the vector of residuals.

The measurement model consists of two equations with the following expressions:

$$X={\Lambda }_{\chi }\xi +\delta$$

(2)

$$Y={\Lambda }_y\eta +\epsilon$$

(3)

In Equation (2), $\xi$ is the exogenous latent variable of order m × 1 (mainly referring to livelihood capital, environmental cognition, and policy cognition), and $\chi$ is the exogenous observed variable of order p × 1; ${\Lambda }_{\chi }$ is the matrix of order p × m, which is the factor loading matrix of the exogenous observed variable $\chi$ on the exogenous latent variable $\xi$; $\chi$ is the exogenous of the observed variable of order p × m; and $\delta$ is the vector of measurement errors of order p × 1.

In Equation (3), $\eta$ the endogenous latent variable of order n × 1 (mainly referring to environmental perception, policy perception, and green production behavior of farmers), $Y$ is the endogenous observed variable of order q × 1, ${\Lambda }_y$ is the matrix of factor loadings of the endogenous observed variable on the endogenous latent variable $\eta$ of order q × n, $Y$ is the endogenous observed variable of order q × n, and $\epsilon$ is the vector of measurement errors of order q × 1.

3.4. Test of Mediating Effects

The mediating effect between farmers’ environmental cognition and policy cognition is assessed using Sobel’s test [33] and the sequential test (Figure 2). The initial stage is to determine the significance of coefficient a1. This suggests that the cognition of producers regarding green production practices does not have a substantial impact on their subsistence capital. If this is the case, the mediating impact is not examined. It is crucial to assess the significance of coefficients b and c in the event that coefficient a1 is significant. It is also imperative to determine the significance of coefficient a2 in the event that coefficients b and c are significant. Farmers’ cognition has a full mediating impact if coefficient a2 is not significant; otherwise, it has a partial mediating effect. Finally, the Sobel test must be repeated if either coefficient b or c is not statistically significant. If the test is significant, it is presumed that risk perception has a mediating effect. If not, it is hypothesized that agrarian perspectives do not have any mediating effect.

4. Analyses and Results

4.1. Reliability and Validity Tests

The reliability and validity of the questionnaire were evaluated using SPSS 21.0 software to ensure the credibility and validity of the study’s findings. Cronbach’s alpha (Cronbach’s coefficient) was calculated for each indicator in the reliability test: natural capital (0.808), physical capital (0.805), human capital (0.735), financial capital (0.848), social capital (0.853), environmental cognition (0.88), policy cognition (0.728), and green production behavior (0.84). The questionnaire’s reliability is indicated by the fact that the Cronbach’s coefficient value for each variable exceeds 0.7 (see Table 5).

The validity of the instrument was assessed using SPSS 21.0 (see

Table 6. KMO and Bartlett’s test.

KMO		0.857
Bartlett’s Sphericity Check	Approximate chi-square	2904.485
	df	190
	Sig.	0.000

). The Bartlett’s sphere test produced an approximate chi-square value of 2973.766, with a significance of Sig = 0.000 (p < 0.01), and the KMO value was 0.857. These findings indicate that the variables in this paper possess a more rigorous structure and are appropriate for additional research.

4.2. Model Testing

By integrating the research hypothesis of this study with the theoretical model illustrated in Figure 1, the first structural equation model of subsistence capital, environmental cognition, and policy cognition was developed using the AMOS 20.0 software. The model includes eight concealed variables in addition to the twenty-three apparent variables. The data in this investigation were in agreement with the fitted values of GFI, AGFI, NFI, CFI, and IFI, which were 0.899, 0.869, 0.863, 0.912, and 0.913, respectively. A model is deemed to have a satisfactory fit if its relative fit index (RMSEA) is less than 0.10. An RMSEA of less than 0.05 indicates a highly exceptional match. However, an RMR of 0.204, which exceeds the threshold of 0.05, suggests a poor fit assessment.

As a result, this paper conducted a preliminary path test on the model’s path and found that the path coefficient of physical capital is −0.044, with a p-value of 0.221 (>0.1). This was the sole form of capital that did not have a significant impact on green production behavior.

The paradigm is modified by this study, which eliminates the physical capital component from the corrective phase. The chi-square degrees of freedom ratio (χ²/df) is 1.717 after rectification, as per the measurement standard’s criteria. The corresponding values for GFI, AGFI, NFI, CFI, and IFI are 0.931, 0.911, 0.911, 0.096, and 0.096, respectively. The RMSEA of the absolute fitness index, post-model modification, is 0.045, which satisfies the model’s criteria despite the fact that the value-added fitness index (RMR) exceeds the model. The model fitting assessment criteria are satisfied by 0.045. The subsequent indicators indicate that the structural equation model’s overall model fit for the factors that influence producers’ green production behavior has improved, as determined by statistical analysis. Following rectification, the coefficients of the route are as shown in Figure 3. Note that → represents the causal relationship between latent variables, which points from the dependent variable to the effect variable. The residuals of the observed variables in both the measurement model and the measurement error of the structural model are represented by e1–e26.

4.3. Path Analysis

Table 7. Path analysis results.

Path			Estimate	S.E.	C.R.	p
Natural Capital	→	Green Production Behavior	0.062	0.047	1.337	0.003 **
Human Capital	→	Green Production Behavior	0.109	0.047	2.35	0.019 **
Financial Capital	→	Green Production Behavior	0.113	0.055	2.048	0.041 **
Social Capital	→	Green Production Behavior	0.17	0.068	2.516	0.012 **
Natural Capital	→	Environmental Cognition	0.193	0.057	3.371	0.012 **
Human Capital	→	Environmental Cognition	0.174	0.056	3.087	0.002 ***
Natural Capital	→	Policy Cognition	0.186	0.068	2.748	0.006 ***
Human Capital	→	Policy Cognition	0.166	0.066	2.528	0.011 **
Environmental Cognition	→	Green Production Behavior	0.345	0.06	5.71	0.014 **
Policy Cognition	→	Green Production Behavior	0.118	0.056	2.098	0.036 **
Environmental Cognition	→	Policy Cognition	0.435	0.078	5.586	0.004 ***

Note: ** and *** represent significant differences at the 0.05 and 0.01 levels, respectively.

demonstrates that hypothesis H1a is validated by the significant influence of natural capital on farmers’ green production behavior (β = 0.062, p = 0.003 < 0.05). The path coefficient of natural capital on farmers’ green production behavior is 0.062, suggesting that natural capital facilitates farmers’ green production practices. Hypotheses H2a and H3a are therefore validated by the fact that natural capital is a significant predictor of farmers’ environmental cognition (β = 0.193, p = 0.012 < 0.05) and farmers’ policy cognition (β = 0.186, p = 0.006 < 0.01). Townships that are situated in close proximity to the county city demonstrate higher output rates per unit area of land, which is indicative of a higher level of farmers’ awareness regarding agricultural investment. Additionally, municipalities that are situated in close proximity to the county capital exhibit a greater level of concern for agricultural-related matters. The adoption of environmentally sustainable agricultural practices is facilitated by the extent of awareness and adherence to these policies by farmers, which is positively correlated with the efficacy of policy implementation. Consequently, hypotheses H1a, H4a, and H5a are substantiated.

Hypothesis H1c is therefore validated by the fact that human capital has a substantial impact on the green production behavior of producers (β = 0.109, p = 0.019 < 0.05). The path coefficient of human capital on farmers’ green production behavior is 0.109, suggesting that the decision-making process of farmers regarding this behavior is positively influenced by the involvement of family labor in agricultural production. Additionally, hypotheses H4c and H5c are confirmed by the fact that human capital is a statistically significant predictor of farmers’ environmental cognition (β = 0.174, p = 0.002 < 0.01) and policy cognition (β = 0.166, p = 0.011 < 0.05). The likelihood of employing green production practices is increased by the ease with which more educated producers can understand and participate in the implementation of policy. As a result, hypotheses H1c, H4c, and H5c are maintained.

Hypothesis H1d is therefore validated by the fact that financial capital has a substantial impact on the green production behavior of producers (β = 0.113, p = 0.041 < 0.05). The path coefficient of financial capital on the green production behavior of producers is 0.113. Farmers’ reserves are bolstered by an increase in their domestic income, which in turn reduces the risks associated with agricultural production investments. Additionally, enhanced access to agricultural financing promotes more effective investment in agricultural materials. As a result, hypothesis H1d is confirmed by the substantial positive impact of financial capital on green production behavior. Nevertheless, hypotheses H4d and H5d are invalidated in this study, as financial capital does not indirectly influence farmers’ cognitive behavior.

Hypothesis H1e is therefore confirmed by the fact that producers’ green production behavior is significantly influenced by social capital (β = 0.17, p = 0.012 < 0.05). The path coefficient indicates that the adoption of agricultural green production technology is facilitated by the increased opportunities for producers to partake in skills training. Furthermore, stronger neighborhood relationships facilitate the dissemination of information and enhance social perceptions. A more robust neighborhood connection fosters the dissemination of knowledge and fosters a positive social perception, suggesting that these relationships enhance farmers’ environmental identity and facilitate communication. The considerable positive impact of social capital on green production behavior supports hypothesis H1e. Nevertheless, hypotheses H4e and H5e are rendered invalid in this study because they do not indirectly influence producers’ green production behavior through cognitive behavior. The path coefficients for producers’ environmental cognition and policy cognition regarding green production behavior are 0.345 and 0.118, respectively, with significance levels that are less than 0.05. This suggests that the adoption of green production practices by producers is facilitated by both environmental and policy cognition. As a result, a greater propensity to engage in green production behaviors is associated with a higher level of cognitive awareness of agricultural production, thereby confirming H2 and H3.

In conclusion, the path analysis revealed that farmers’ green production behavior is significantly positively influenced by all components of livelihood capital, with the exception of physical capital. Hypotheses H1a, H1c, H1d, and H1e were all verified; moreover, hypotheses H2 and H3 were validated, and producers’ green production behavior is positively influenced by both environmental cognition and policy cognition. Nevertheless, an examination of the impact of livelihood capital on environmental cognition and policy cognition demonstrated that only the hypotheses concerning human and natural capital were confirmed, while the remaining capital hypotheses were not (refer to

Table 8. Summary of study hypothesis testing results.

Hypotheses	Hypothetical Content	Results
H1a	Natural capital has a significant positive effect on the green production behavior of farmers	Passed
H1b	Material capital has a significant positive effect on the green production behavior of farmers	Failed
H1c	Human capital has a significant positive effect on the green production behavior of farmers	Passed
H1d	Financial capital has a significant positive effect on the green production behavior of farmers	Passed
H1e	Social capital has a significant positive effect on the green production behavior of farmers	Passed
H2	Environmental cognition has a significant positive effect on farmers’ green production behavior	Passed
H3	Policy cognition has a significant positive effect on farmers’ green production behavior	Passed
H4a	Natural capital has a significant positive effect on environmental cognition	Passed
H4b	Material capital has a significant positive effect on environmental cognition	Failed
H4c	Human capital has a significant positive effect on environmental cognition	Passed
H4d	Financial capital has a significant positive effect on environmental cognition	Failed
H4e	Social capital has a significant positive effect on environmental cognition	Failed
H5a	Natural capital has a significant positive effect on policy cognition	Pass
H5b	Material capital has a significant positive effect on policy cognition	Failed
H5c	Human capital has a significant positive effect on policy cognition	Pass
H5d	Financial capital has a significant positive effect on policy cognition	Failed
H5e	Social capital has a significant positive effect on policy cognition	Failed
H6a	Environmental cognition has a mediating effect on the relationship	Pass
H6b	between farmers’ natural capital and farmers’ green production behavior	Failed
H6c	Environmental cognition has a mediating effect on the relationship	Pass
H6d	between farmers’ material capital and farmers’ green production behavior	Failed
H6e	Environmental cognition has a mediating effect on the relationship	Failed
H7a	between farmers’ human capital and farmers’ green production behavior	Pass
H7b	Environmental cognition has a mediating effect on the relationship	Failed
H7c	between farmers’ financial capital and farmers’ green production behavior	Pass
H7d	Environmental cognition has a mediating effect on the relationship	Failed
H7e	between farmers’ social capital and farmers’ green production behavior	Failed

). As a result, the ensuing mediating effect analysis only investigated the mediating effects of natural capital and human capital.

4.4. Tests of Mediating Effects

4.4.1. Mediation Test of Environmental Cognition in Livelihood Capital and Green Production Behavior

Table 9. Test of the mediating role of environmental cognition on farmers’ livelihood capital and farmers’ green production behavior.

Model	Livelihood Capital → Environmental Cognition	Livelihood Capital → Green Production Behavior	Livelihood Capital → Environmental Cognition → Green Production Behavior
Independent variable	M1	M2	M3
Natural capital	0.211 **	0.123 **	0.050
Mediating variables
Environmental cognition	-	-	0.344 **
R2	0.043	0.022	0.198
F	16.12 **	8.287 **	44.441 **
Independent variable	N1	N2	N3
Human capital	0.181 **	0.176 **	0.116 **
Mediating variables
Environmental cognition	-	-	0.330 **
R2	0.038	0.055	0.217
F	14.259 **	21.142 **	50.053 **

Note: ** represent signifcant diferences at the 0.05 levels, respectively.

demonstrates that environmental cognition (M1, β = 0.211, p < 0.05) and green production behavior (M2, β = 0.123, p < 0.05) are both positively influenced by natural capital. When both environmental cognition and natural capital are considered as independent variables simultaneously, the influence of environmental cognition remains substantial, while that of natural capital becomes negligible. This implies the existence of a mediating role for environmental cognition, which completely mediates the relationship between natural capital and green production behavior. Hypothesis H6a is subsequently validated by the indirect effect of environmental cognition in the trajectory from natural capital to green production behavior of producers (M3, natural capital → environmental cognition → green production behavior) as 0.193 × 0.345 = 0.070.

Human capital has a positive and substantial influence on environmental cognition (N1, β = 0.181, p < 0.05) and green production behavior (N2, β = 0.176, p < 0.05). This indicates that human capital and environmental cognition operate as separate factors in the decision-making processes of producers. Policy cognition has a significant impact, although human capital remains a crucial factor. The findings suggest that environmental cognition serves as a partial mediator between human capital and environmentally sustainable production behavior. The indirect impact of environmental cognition from human capital to farmers’ green production behavior (N3, human capital → environmental cognition → green production behavior) is computed as 0.174 × 0.345 = 0.060, yielding a total effect of 0.079, thereby supporting hypothesis H6c.

4.4.2. Mediation Test of Policy Cognition Between Livelihood Capital and Green Production Behavior

Table 10. Tests of the mediating role of policy cognition on farmers’ livelihood capital and farmers’ green production behavior.

Model	Livelihood Capital → Policy Cognition	Livelihood Capital → Green Production Behavior	Livelihood Capital → Policy Cognition → Green Production Behavior
Independent variable	M1′	M2′	M3′
Natural capital	0.228 **	0.123 **	0.072
Mediating variables
Policy cognition			0.221 **
R2	0.053	0.022	0.09
F	20.106 **	8.287 **	17.928 **
Independent variable	N1′	N2′	N3′
Human capital	0.199 **	0.176 **	0.135 **
Mediating variables
Policy cognition	-	-	0.205 **
R2	0.048	0.055	0.114
F	18.392 **	21.142 **	23.205 **

Note: ** represent signifcant diferences at the 0.05 levels, respectively.

indicates that natural capital positively influences policy cognition (M1′, β = 0.228, p < 0.05) and green production behavior (M2′, β = 0.123, p < 0.05). The influence of policy cognition remains substantial when both natural capital and policy cognition are treated as independent variables, but the influence of natural capital becomes negligible. The findings suggest that policy cognition serves as an overarching mediator between green production practices and natural capital. The indirect effect of policy cognition on natural capital concerning farmers’ green production behavior (M3′, natural capital → policy cognition → green production behavior) is computed as 0.186 × 0.118 = 0.022, thereby substantiating hypothesis H7a.

Human capital positively influences policy cognition (N1′, β = 0.199, p < 0.05) and green production behavior (N2′, β = 0.176, p < 0.05). The influence of policy cognition is significant when both human capital and policy cognition are treated as independent variables, whereas the effect of human capital continues to be considerable. Policy cognition serves as a partial mediator between producers’ green production behavior and livelihood capital. The indirect effect of policy cognition on farmers’ green production behavior via human capital (N3′, human capital → policy cognition → green production behavior) is computed as 0.166 × 0.118 = 0.020, while the total effect is 0.039, thereby confirming hypothesis H7c.

The aforementioned analysis suggests that physical capital does not have a significant impact on farmers’ green production behavior. Consequently, the hypotheses H6b and H7b are invalid, as there is no mediating effect of cognitive behavior between physical capital and farmers’ green production behavior. Additionally, hypotheses H6d, H7d, H6e, and H7e are rendered invalid due to the fact that financial capital and social capital have a direct impact on the green production behavior of farmers, but they do not indirectly influence it through cognitive behavior. The results imply that the adoption of green production practices is more significantly influenced by farmers’ cognitive behavior than by livelihood capital. This suggests that farmers should evaluate the value of the benefits derived from investing in production inputs when making business decisions. As a result, it is imperative to prioritize the facilitation of the transition of farmers from awareness to actionable behavior in the promotion of environmental conservation, as the indirect influence of livelihood capital on farmers’ environmentally sustainable production practices exceeds the direct influence.

5. Discussion

The aim of this study is to investigate how policy cognition, environmental cognition, and livelihood capital affect producer green production behavior. Except for physical capital, every capital in farmers’ livelihood capital has been proven to greatly affect their green production methods. Environmental and policy cognition have more influence than livelihood capital and may help to effectively promote the positive benefits of natural and human capital on producers’ green production practices. This result satisfactorily answers the research questions.

An important component of China’s agricultural modernization and a clear example of the idea of green development is green production in agriculture [34]. As the main emphasis of agricultural production and operation, a range of factors, including cognitive ability and financial endowment, might affect and constrain producers’ green production cognition and behavior [35,36]. This work revealed the phenomena of “low behavior” and “high cognition” in the green output of vegetable growers. This is consistent with the findings that farmers’ knowledge of green production differs significantly from their behavior [37]. Thus, we must give the promotion of environmentally friendly agricultural techniques among farmers first priority if we are to attain “knowledge and action”.

This article revealed that livelihood capital, natural capital, human capital, financial capital, and social capital may considerably affect the green production behavior of producers. This is in line with the results of other researchers [38,39] who have come to the conclusion that social capital, human capital, and natural capital help producers in the green production sector to act. Geng Strait [38] also found that physical capital limits the green production behavior of producers. Nevertheless, the study presented in this article shows that the green production behavior of producers is not much influenced by physical capital. This is true because farmers in the research region had a tiny per capita arable land area, a low automation rate of vegetable farming, a mostly manual sowing technique, and quite similar farm equipment and tools. Consequently, the impact on the ecological production methods of producers is minimal. This study’s empirical data also showed that capital endowment affects producers’ green production behavior at many different cognitive levels. Environmental cognition [40] has the most major impact on producers’ green output. Moreover, government incentives, policy and legal restrictions, and environmental crisis awareness help farmers use organic fertilizer in a better way [35].

This paper has some limitations. It is a cross-sectional study, which measured producers’ green production behavior, cognition, and livelihood capital at a specific timepoint; therefore, the results of the research show the static connection between the variables. Subsequent studies will make use of monitoring survey data to represent the dynamic changes in producers’ green production behaviors. Second, although it is based on the livelihood sustainability analysis framework and involves an impact evaluation of green production behavior from the viewpoint of several capitals, this study neglected to consider the inter-relationships between different capitals and their effects on farmers’ green production behavior. This is both a constraint and a possible focus for further studies.

6. Research Findings and Suggestions

The survey data of 364 vegetable farmers in the primary production area of Northwest Inland Export Vegetables are employed in this paper to examine the factors that influence the green production behavior of farmers. The following deductions are made:

(1)

Among vegetable farmers, the average values for the use of organic fertilizer, green prevention and control, scientific use of medicine, pesticide and chemical fertilizer waste disposal, and scientific use of medicine are 3.19, 3.11, 3.21, and 3.23, respectively. The evaluations of adoption behavior for green preventive and control systems are very poor. Farmers score 3.36 on policy cognition and 3.27 on environmental cognition. This helps to clarify ideas of “low behavior” and “high cognition”.

(2)

Although it is not clear how physical capital affects farmers’ green production methods, natural, human, financial, and social capital may greatly improve such practices.

(3)

Natural and human capital indirectly influence farmers’ sustainable production methods by means of environmental and policy awareness. While operating as a partial intermediate between human capital and green production behavior, environmental cognition and policy cognition work as full mediators between natural capital and green production behavior. Farmers’ green production methods are more influenced by their cognitive habits than by the impact of livelihood capital.

Based on the above research findings, the following policy implications are drawn:

(1)

Enhance farmers’ understanding of sustainable production policies and environmental education. To enhance farmers’ awareness of the rural ecological environment and agricultural product safety and to cultivate diverse ecological and policy cognitions regarding self-interest, altruism, and ecological values, it is essential to utilize new media to actively inform farmers about soil stagnation, nutrient depletion, and rural environmental pollution resulting from excessive fertilizer use, as well as food safety concerns stemming from using banned pesticides or elevated pesticide residues.

(2)

Enhance agricultural subsidies for sustainable production. To promote the shift in farmers’ mindset towards sustainable production practices, it is imperative to enhance subsidies for green production through diverse avenues, such as the provision of organic fertilizers, organic pesticides, insecticidal lamps, and nets; complimentary training on various sustainable production techniques; guidance on the scientific application of organic inputs; and encouragement of physical pest management strategies while minimizing the use of fertilizers and pesticides during production.

(3)

Improve farmers’ livelihood capital via many approaches. Natural capital includes facilitating land transfers and extensive operations; human capital includes improving grassroots training in sustainable agricultural production technologies and advancing farmers’ expertise in sustainable practices; actively encouraging farmers to participate in cooperative organizations and broadening social resource information channels by fostering mutual trust among farmers, village collectives, and promoting opportunities for interaction and exchange; progressively enhancing the use of agricultural machinery and equipment and elevating the standards of sustainable agricultural practices. Small farmers should be encouraged to use agricultural production trusteeship services or the policy of subsidizing agricultural machinery purchases to solve shortages of agricultural equipment and other physical capital. Using green subsidies, strongly support organic fertilizers, biopesticides, and both physical and chemical prophylactic and control techniques. Concurrently, strengthen financial institutions connected to agriculture to enhance their policies, thus benefiting and supporting the industry.