Ecological Awareness, Policy Perception, and Green Production Behaviors of Farmers Living in or near Protected Areas

Abstract

Research highlights: Ecological policies must balance ecosystem protection by promoting the sustainable livelihoods of farmers living in or near protected areas; however, the intrinsic motivations of farmers to adopt green production behaviors (GPBs) are poorly understood. Background and objectives: We explored how ecological policies affect the GPBs of farmers in agroforestry. Materials and methods: We conducted questionnaires of farmers in 11 counties of Sichuan Province, China, with abundant protected areas and large-scale agroforestry, after which a structural equation model of farmers’ ecological awareness, policy perception, and GPBs was constructed. Results: (1) Ecological policies can stimulate farmers’ GPBs by improving their ecological awareness, creating positive subjective norms, and inducing the “herd effect”. Increases in protection intensity and scope amplify the pressures on farmers to maintain more than long-term policy consistency. (2) Green production is more time-consuming, laborious, expensive, and difficult to learn compared with traditional production methods, which have somewhat limited GPBs adoption. (3) In the rural “acquaintance society”, information and communication from others have a substantial impact on farmers’ perceptions, attitudes, and behaviors; thus, positive subjective norms from formal and informal channels could promote GPB adoption. Conclusions: Future policies should prioritize environmental education over environmental publicity by helping farmers understand the long-term relationship between ecological protection and economic development, teaching individual environmental responsibility, enhancing positive feedback to farmers who adopt GPBs, actively exploring mechanisms for realizing the value of ecological products, and improving farmers’ management skills and learning ability.

1. Introduction

The conflict caused by the high spatial overlap between ecologically protected and relatively undeveloped rural areas is particularly serious in developing countries [1]. Ecological policies aim to protect the authenticity and integrity of ecosystems but must also promote the sustainable livelihoods of farmers who have lived in the vicinity of protected areas for many generations [2,3]. The protected areas in China include various types, such as national parks, nature reserves, and forest parks—especially the establishment of national parks, which has further expanded the area of protected areas. Due to historical and practical reasons, many rural residential areas are surrounded by these protected areas, as is much of the forest and farmland they own. Therefore, rural communities and protected areas are spatially connected, adjacent, or even partially overlapping. A series of ecological protection policies were implemented, mainly including the establishment of national parks, projects aiming to return farmland into forests, and a number of cutting promotion and restriction policies. These policies have an impact on farmers’ production activities; that is, as the area of protected areas further increases, including non-commercial and economic forests, farmers’ production activities become more focused on agroforestry. Agroforestry is a typical sustainable production practice that delivers excellent ecosystem services to improve the output and quality of agricultural and forestry products and increase the added value of products without damaging the original ecosystem service function [4,5] while also enhancing and stabilizing rural livelihoods [6,7]. Agroforestry activities should be implemented under the constraint of ecological protection; however, few farmers consider whether their production processes are environmentally friendly [8]. Moreover, the structure and function of agroforestry are influenced by many factors, and the impact of agroforestry on ecosystems remains controversial. For example, inappropriate management methods may lead to negative environmental consequences such as water scarcity, nutrient deficiency, increased disease and pest infestations, excessive interference with forest land and wildlife habitats, and resource waste [9,10,11,12]. These behaviors pose a significant threat to the stability, sustainability, and diversity of the ecosystem around protected areas [13]. Therefore, it is necessary to encourage farmers to adopt green production behaviors (GPBs).

Despite several studies on green consumption behaviors in society, little research has been conducted on GPBs. The GPBs of local farmers involve a positive interaction between humans and nature, including all behaviors that consciously protect the ecosystem, save energy and resources, and reduce environmental pollution in agricultural production and non-timber forest operation [14,15]. The connotation of GPBs is similar to that of pro-environmental [16] and green consumption behaviors [17] while emphasizing that such behaviors should expand to agricultural and forestry production activities. Farmers adopting GPBs exhibit positive externality [18], which is beneficial for ecological protection but may also experience higher private costs. As such, if farmers cannot benefit from ecological protection, they may lose interest. Therefore, policymakers should consider the factors influencing farmers’ GPB adoption and propose ecological policies that effectively encourage GPBs.

According to the theory of “rational farmers,” farmers’ decision-making processes are dominated by the economic benefits and risks [14,19]. When the economic benefits of adopting GPBs are higher than those of traditional production methods, farmers are likely to consciously and actively adopt such behaviors [20]. Moreover, monetary incentives can significantly improve their adoption [21]. However, monetary subsidies are not unlimited and cannot fully cover the time, money, and output loss that farmers experience when adopting GPBs. Therefore, it is crucial to identify the intrinsic motivations for farmers to adopt such behaviors. Previous research on this topic has identified three main factors. First, most local farmers maintain their traditional ideology and exhibit resistance to change, that is, the status quo bias, whereby people systematically prefer to keep their current practices because they perceive any change as a loss [22]. Second, farmers’ ecological awareness modifies their desire to pursue maximum economic benefits [15,23]; that is, farmers with high awareness and a better understanding of the negative impact of ecological damage are more likely to take environmentally friendly measures [24]. Finally, subjective normative pressure from society and other people may have a latent impact on farmers’ awareness and behaviors [25]. However, few studies have explored the mechanism and influencing pathways of and between these factors.

Furthermore, most research is focused on GPBs in agricultural activities but rarely include the constraints of protected areas. Farmers typically own a small amount of arable and a large amount of forest land; therefore, their agroforestry activities have an important impact on protected areas. As an external incentive, ecological policies affect farmers’ GPBs both directly and indirectly by changing their awareness and social norms [14,25]. Farmers may exhibit adaptive or resistant behavior based on internal and external conditions caused by various ecological policies. For example, farmers are not typically sufficiently motivated to cooperate with the government to protect the ecosystem; however, high policy recognition can increase the probability of farmers adopting GPBs [2]. Moreover, farmers’ ecological values are constantly adjusted in alignment with the interaction between the deterioration of the ecological environment and the implementation of ecological policies [8].

In this study, we employ the structural equation model (SEM) to explore the key factors influencing the GPBs of farmers’ agroforestry activities in and around protected areas as well as the mechanisms by which ecological policies impact their GPBs. Based on this goal, we attempt to (1) evaluate the effects of farmers’ protective attitude, awareness, subjective norms, and perceived behavioral control on their GPBs, (2) clarify the effect of the observation variables on each latent variable, and (3) measure the indirect interactions between these latent variables. Finally, we propose policy suggestions to help coordinate the ecological protection and economic development of land surrounding protected areas.

2. Theoretical Framework and Hypotheses

The theory of planned behavior is typically used to determine the relationship among farmers’ attitudes, perceptions, and behaviors. It states that intentions to perform behaviors of different kinds can be predicted with high accuracy from attitudes toward the behavior, subjective norms, and perceived behavioral control; these intentions, together with the perception of behavioral control, account for considerable variance in actual behavior [26]. According to this theory, if farmers experience higher private costs when adopting GPBs in agroforestry, they have a higher probability of adopting them. Additionally, the adoption of GPBs is suggested to be primarily affected by their attitudes, subjective norms, and perceptual behavior control, and studies have shown that these factors are conducive to an individual adopting pro-environmental behaviors [27,28,29]. However, few studies have focused on the agroforestry activities of farmers living in protected areas. Therefore, we constructed a theoretical framework according to the theory of planned behavior and proposed the following hypotheses.

First, an attitude refers to the positive or negative feelings held by individuals [26]. In addition, ecological policies typically place publicity and the education of surrounding communities in a highly important position. If these efforts are effective, farmers’ ecological awareness should have a positive impact on their GPBs. The greater the perception of ecological policies, the greater the positive impact. Thus, the following hypotheses are proposed:

Hypothesis 1a. 

Farmers’ ecological awareness has a positive impact on their GPBs.

Hypothesis 1b. 

Farmers’ perception of the effects of ecological policies has a positive impact on their GPBs.

Subjective norms refer to the social pressures felt by farmers to implement a specific behavior [26], that is, what others think they should do places a certain group of pressure on farmers, and the intensity of their subjective desire to maintain consistency with others’ opinions impact their behavior. Ecological policies affect not only farmers themselves but also other people in the same situation, leading to the “herd effect”. Thus, a positive influence exerted by others should encourage farmers to adopt GPBs [30], and the following hypothesis is proposed:

Hypothesis 1c. 

Farmers’ subjective normative perception has a positive impact on their GPBs.

Perceived behavioral control refers to the perceived degree of difficulty required to adopt a certain behavior [26]. Most GPBs differ from the behavior farmers have become accustomed to. Rational farmers measure their own capabilities; that is, the more resources and opportunities and the fewer expected obstacles they perceive, indicating higher control of more external factors, the more likely farmers are to adopt GPBs. Thus, the following hypothesis is proposed:

Hypothesis 1d. 

The perceived behavioral control of farmers has a positive impact on the adoption of GPBs.

Finally, a reciprocal influence exists among attitude, subjective norms, and perceived behavior control. Thus, the following hypotheses are proposed:

Hypothesis 2. 

Farmers’ perception of the effects of ecological policies has a positive impact on their ecological awareness.

Hypothesis 3a. 

Farmers’ subjective normative perception has a positive impact on their ecological awareness;

Hypothesis 3b. 

Farmers’ subjective normative perception has a positive impact on their perception of the effects of ecological policies.

Hypothesis 4a. 

The perceived behavioral control of farmers has an important impact on their ecological awareness;

Hypothesis 4b. 

The perceived behavioral control of farmers has an important impact on their perception of the effects of ecological policies.

Therefore, we constructed nine hypotheses to describe the relationship between five types of latent variables (Figure 1), including farmers’ GPBs, ecological awareness (EA), perception of the effects of ecological policies (PE), subjective norms (SN), and perceived behavior control (PB).

3. Data and Methods

3.1. Study Area

Sichuan Province is a large mountainous province in China with rich forest resources, an important ecologically strategic location, and a fragile ecological environment. According to data from the official website of the Sichuan Provincial Government, the province contains several types of protected areas, covering 23.3% of its entire land area. As the main habitat of the giant panda (Ailuropoda melanoleuca), Sichuan Province encompasses the majority of the Giant Panda National Park. Owing to its rich forest resources, various modes of agroforestry have recently been implemented in villages around nature reserves and the Giant Panda National Park. These local rural villages are characterized by poor economic development, with a per capita disposable income of only 18,672 yuan as of 2022. As such, Sichuan Province is a typical and representative region suitable for research into farmers’ GPBs in agroforestry. Therefore, we selected 11 counties of Sichuan Province with abundant protected areas and large-scale agroforestry as the study area (Figure 2).

3.2. Sample and Data

In this study, we used questionnaire data to analyze farmers’ GPBs. A trained research team traveled to the study sites between March to August 2019 to conduct one-to-one structured questionnaire surveys. The questionnaire included the following main topics: basic demographics of farmers and their families; farmers’ land, assets, and social capital endowment; and farmers’ ecological awareness, protection attitudes, and evaluation of ecological policies. Small-scale pilot surveys were conducted to verify the completeness and effectiveness of the questionnaire. After eliminating invalid samples, 975 responses from 79 villages were collected, with 10–30 households surveyed in each village according to the population scale. Most survey samples were collected from Baoxing County, Pingwu County, and Dujiangyan County-level cities located in the south, north, and middle of the Giant Panda National Park, respectively, because of their abundant protected areas and dense population (Figure 3).

3.3. Methods

Structural equation modeling (SEM) is a useful method when measuring latent variables is challenging [31]. SEM establishes, estimates, and tests the causality model. Additionally, compared with traditional methods, SEM not only tests the relationship between observation, latent, and interference or error variables but also integrates factor and path analyses, which can obtain the direct, indirect, or total effects of independent variables on dependent variables. In this study, the farmers’ subjective perception and behavioral intention were latent variables difficult to directly observe and accurately measure, but they could be characterized by a series of observation variables, making them suitable for SEM analysis. SEM comprises two basic models: the measurement and the structural model.

The measurement model includes the latent and observation variables, and the equations are as follows:

$$\mathrm{Y}=\mathsf{\lambda }\mathsf{\eta }+\mathsf{\epsilon }$$

$$\mathrm{X}=\mathsf{\lambda }\mathsf{\phi }+\mathsf{\delta }$$

where $\mathrm{Y}$ and $\mathrm{X}$ are internal and external variables, respectively; $\mathsf{\lambda }$ is the regression type; η and $\mathsf{\phi }$ are vectors composed of endogenous and exogenous latent variables, respectively; and $\mathsf{\epsilon }$ and $\mathsf{\delta }$ denote the variance/covariance type.

The structural model is an explanation of the causal relationship between potential variables, the equation of which can be written as:

$$\mathsf{\eta }=\mathsf{\gamma }\mathsf{\phi }+\mathsf{\beta }\mathsf{\eta }+\mathsf{\epsilon }$$

where $\mathsf{\phi }$ and $\mathsf{\eta }$ are the vector types, $\mathsf{\gamma }$ and $\mathsf{\beta }$ are the regression types.

We included five latent variables in this study, and the observation variables for each latent variable are shown in

Table 1. Description of variables used in the structural equation modeling (SEM).

	Variable Name	Mean	Std.	Factor Load	Cronbach’s Alpha	KMO
EA	Ecological awareness	3.297	-		0.862	0.860
EA1	A good ecological environment is conducive to physical and mental health	3.490	0.924	0.774
EA2	A good ecological environment is conducive to economic development	2.916	1.164	0.872
EA3	Human destruction of nature will have disastrous consequences	3.464	0.908	0.863
EA4	Wild animals and plants have the same right to life as human beings	3.275	0.969	0.803
EA5	Everyone should protect the ecological environment	3.339	0.912	0.696
PE	Perception of the effects of ecological policies	3.621	-		0.903	0.865
PE1	The construction of nature reserves has played a beneficial role in ecological protection	3.689	1.266	0.794
PE2	Various ecological projects (returning land for farming to forestry, natural forest protection, etc.) have played a beneficial role in ecological protection	3.516	1.347	0.800
PE3	Timber-cutting prohibition/restriction policies have played a beneficial role in ecological protection	3.667	1.252	0.898
PE4	Expansion of protected areas has played a beneficial role in ecological protection	3.177	1.429	0.870
PE5	Ecological migration has played a beneficial role in ecological protection	4.054	1.173	0.912
SN	Subjective normative perception	3.526			0.830	0.785
SN1	I understand that damage to the environment and resources will be punished	3.777	1.219	0.752
SN2	I agree with the green protection concept promoted by the government	3.659	1.206	0.803
SN3	I think active ecological protection will be respected by society	3.229	1.306	0.752
SN4	My relatives and friends often communicate about ecological and environmental issues	3.496	1.238	0.802
SN5	My relatives and friends have implemented green production practices	3.470	1.309	0.750
PB	Perceived behavioral control	2.958			0.877	0.850
PB1	I have a good understanding of green production	3.145	1.490	0.847
PB2	I have a good understanding of the laws and regulations related to ecological protection	2.916	1.448	0.897
PB3	I think my family has a strong ability to improve our livelihood level	2.980	1.478	0.795
PB4	I think my learning ability is good	2.920	1.443	0.790
PB5	The environment and public infrastructure in the village are very helpful to me	2.831	1.518	0.766
GPBs	Green production behaviors (GPBs)	3.151			0.844	0.828
GPB1	I have invested greater production costs to adopt GPBs	3.542	1.264	0.706
GPB2	I have given up a portion of my economic earnings to adopt GPBs	3.022	1.347	0.872
GPB3	I have invested more working hours to adopt GPBs	2.857	1.337	0.859
GPB4	I have invested a larger labor force to adopt GPBs	2.699	1.362	0.803
GPB5	I have mastered new green production technologies to adopt GPBs	3.635	1.327	0.676

. EA mainly examined farmers’ understanding of the importance of the ecological environment and the relationship between people and the ecological environment. PE not only included the perception of ecological protection policies and key projects that have been implemented for a long time in the local area, but also the perception of the expansion of protected areas and ecological migration policies due to the establishment of national parks. SE mainly included mandatory and descriptive norms. PB mainly considered farmers’ perceived control over green production knowledge, livelihood and learning abilities, and infrastructure construction. In terms of GPBs, local physical and geographical conditions determined that farmers had less arable and more forest land. Additionally, agroforestry is the main production activity in forest land, which has a greater impact on the ecosystem. Therefore, it is necessary to study the GPBs of agroforestry farmers. These include a series of behaviors that are beneficial for ecological protection during forest clearing, selection of planting varieties and densities, irrigation, fertilization and pesticide application, waste disposal, growth of understory livestock and poultry, fecal pollution, and diseased and dead animal disposal. Notably, we cannot measure the individual willingness to adopt these behaviors. Additionally, it must be considered that no matter which behavior is adopted, there will be a certain private cost associated with it. Based on this, we ensured that farmers understood the connotation of GPBs and characterized their willingness to adopt them from the perspective of loss of income, increased costs, increased labor time, increased labor force, and technological improvement. Each observation variable was divided into five levels according to the Likert scale method: 1. strongly disagree, 2. disagree, 3. neither agree nor disagree, 4. agree, and 5. strongly agree, which indicates the degree to which farmers relate to each variable. According to the average score and standard deviation of the indicators, farmers were greatly affected by various previous and recent ecological policies. Their ecological awareness was at an intermediate level, their subjective norms were relatively strong, and their awareness and ability to control relevant resources were weak. Additionally, farmers were more likely to adopt GPBs if they could invest in production costs and technology improvements rather than increasing working hours or labor force.

3.4. Statistical Analysis of the Model

IBM SPSS v.24.0 and AMOS v.20.0 were used to conduct exploratory factor analyses of reliability and validity (Table 1). The factor loads for all indicators were above 0.6, confirming the rationality of the indicator system. Cronbach’s Alpha index, which is used to test reliability, was higher than 0.8. The Kaiser-Meyer-Olkin (KMO) value, used to test validity, was higher than 0.7. Bartlett’s spherical test proved that the indicator system was reliable and effective. We set the critical value of the modification indices in the initial hypothesis model to 20, then altered the indicators with higher modification indices to ensure that all modification indices were less than 20. Three indicator systems (absolute, incremental, and reduced fitness statistics) were used to test the model’s fitness; the results showed that the model had high consistency and good fitness (

Table 2. Model fitness test index and evaluation standard.

Fitness Indicator		Optimal Criterion	Actual Value
Absolute fitness statistics	Square root of approximation error (RMSEA)	<0.05	0.050
	Goodness of fit index (GFI)	>0.90	0.912
Incremental fitness statistics	Modified goodness of fit index (AGFI)	>0.90	0.893
	Regulatory fit Index (NFI)	>0.90	0.905
	Comparison fit index (CFI)	>0.90	0.945
	Value-added adaptation index (IFI)	>0.90	0.945
	Tacker-Lewis index (TLI)	>0.90	0.937
Reduced fitness statistics	Parsimony goodness-of-fit index (PGFI)	>0.50	0.747
	Parsimony-adjusted CFI (PCFI)	>0.50	0.837
	Parsimony-adjusted NFI (PNFI)	>0.50	0.803

).

4. Results and Discussion

4.1. Personal and Family Characteristics

The basic personal and family characteristics of the surveyed farmers are shown in

Table 3. Basic demographics of the rural households sampled in this study.

Variable	Variable Name	Assignment Statement	Average Value	Minimum Value	Maximum Value	Standard Deviation
basic characteristics of the household head	gender	1 = male; 2 = female	1.13	1	2	0.34
	age	years	47.55	19	93	12.51
	marital status	1 = married; 2 = unmarried	1.31	1	2	0.46
	education level	how many years did he/she go to school	5.41	0	21	2.30
	health condition	1 = good; 2 = moderate; 3 = bad	1.42	1	3	0.65
	type of job	1 = only engaged in agriculture/forestry; 2 = part-time farmer; 3 = non-agricultural employment	1.35	1	3	0.60
basic household characteristics	family population	total number of households	4.31	1	14	1.70
	annual family income	yuan	61,608.90	1100	796,810	78,254.50
	residence in or surrounded by the national park	1 = yes; 2 = no	1.78	1	2	0.41
land resource endowment	total area of cultivated land	mu	5.06	0	137	8.44
	cultivated land quality	1 = good; 2 = moderate; 3 = bad	2.13	1	3	0.64
	average distance between cultivated land and their home	meters	1070.84	10	10,500	1284.66
	total area of forest land	mu	23.66	0	50,000	62.39
	forest land quality	1 = good; 2 = moderate; 3 = bad	2.06	1	3	0.57
	average distance between forest land and their home	meters	2154.59	40	50,000	3481.40

. The majority of the household heads are male, married, and elderly, with a low education and average health level; most of them engaged in agricultural and forestry activities and lived in or were surrounded by the Giant Panda National Park. Despite having many members in each rural household, the average annual income of rural households was low. Furthermore, on average, the endowment of forest land resources was better than that of cultivated land resources because the local terrain was predominantly mountainous with steep slopes. Thus, agroforestry was a common choice for local farmers throughout the study area.

4.2. Measurement Model Analysis

In the SEM, the influence coefficient of the measurement model indicates the extent to which the observation variable can reflect the level of the latent variable. The observed variable is the “effect”, and the latent variable is the “cause” [31]. According to the measurement model analysis results (Figure 4), we present the following major findings.

(1)

The two indicators with a high impact on farmers’ ecological awareness were “a good ecological environment is conducive to economic development” (0.86) and “human’s destruction of nature will have disastrous consequences” (0.82). Most farmers hold ecological value from a personal perspective, but only those with high ecological awareness could better understand the benefits of a healthy environment and the serious consequences of ecological damage from a macro and long-term perspective. According to the results, farmers’ awareness of biodiversity conservation has significantly reduced. As ecological policies have been implemented over the years, the number of wild animals has increased rapidly, and human–wildlife conflicts have intensified. If these cannot be alleviated, this may become a hindering factor regarding farmers’ adoption of GPBs, which are conducive to biodiversity conservation.

(2)

The three indicators that had a high impact on farmers’ perception of the effects of ecological policies were “ecological migration (0.91)”, “timber cutting prohibition/restriction policy (0.88)”, and “expansion of protected areas (0.86)”. These policies are beneficial for protecting the integrity and authenticity of the ecosystem but can force farmers to leave their homeland and change their traditional livelihood and lifestyle. Farmers with a more positive attitude may exhibit higher support for these policies and a higher likelihood of adopting GPBs. Correspondingly, the construction of protected areas and several ecological projects have been implemented stably and sustainably for several decades, with farmers consenting to continue production activities under the constraint of these policies.

(3)

The subjective norm perception of farmers was affected by various indicators with a low coefficient. Normative pressure from the outside world is multifaceted, so farmers’ subjective norm perception is also complex and interrelated. The indicator with the highest impact was “my relatives and friends often communicate about ecological environmental issues” (0.76). Despite the many laws and regulations on ecological protection, it is difficult to impose substantive constraints on farmers because of the high supervision costs and poor enforceability. As a result, mandatory norms have less impact on farmers’ protection attitudes than descriptive norms. Although formal channels, such as publicity from the government, village cadres, members of the Communist Party of China, or other advanced individuals, have played a role in promoting farmers’ perception, information transmission and descriptive norms related to informal channels such as communication from relatives and friends are more effective. This is because China’s rural population is an “acquaintance society” with strong geographical relationships. Therefore, the exchange of ecological concepts can effectively enhance farmers’ perception of the subjective norms of the outside world. Indeed, some farmers may adopt GPBs to gain social respect if they genuinely believe it will earn them respect from society and others.

(4)

Two indicators that had a relatively high impact on farmers’ perceived behavioral control were “I have a good understanding of laws and regulations related to ecological protection” (0.91) and “I have a good understanding of green production” (0.83). Most farmers are old, poorly educated, and lack employment skills and learning abilities; therefore, they typically have a weak perception of their own ability and cannot fully understand laws, regulations, and green management information. In addition, rural traffic in mountainous areas is inconvenient, and the environmental infrastructure is poor, which leads to a weak perception of behavioral control related to external factors. To encourage farmers to adopt GPBs, policymakers should not only help farmers improve their own capabilities but also improve necessary objective conditions and resources for green production in agroforestry.

(5)

Three indicators with a high impact on farmers’ GPBs were “I have given up a portion of my economic earnings to adopt GPBs (0.85)”, “I have invested more working hours to adopt GPBs (0.82)”, and “I have invested a larger labor force to adopt GPBs (0.75)”, whereas the adoption of green production technology and greater production costs were not strongly related to their willingness to adopt GPBs. These results show that farmers are more likely to accept a loss of green production than actively change their production mode. In the current circumstances, post-production compensation may better enhance farmers’ GPBs than guidance and supervision during the pre- and mid-production process, although the latter may be a more effective method for reducing ecological damage.

Figure 4. Standardized path coefficients of SEM.

Figure 4. Standardized path coefficients of SEM.

4.3. Path Analysis Results

We used IBM SPSS and Amos v.20.0 to estimate the SEM parameters; the path analysis results are shown in Table 4. H1a, H1c, H3a, H3b, and H4a were verified at a significance level of 1%, and H2a was verified at a significance level of 5%. H1b was reverse verified at a significance level of 5%. Thus, the main results are as follows.

(1)

Farmers’ decision-making was significantly affected by their cognitive attitude. Improving farmers’ ecological awareness and making them realize that healthy ecosystems are an important prerequisite and guarantee for green production will generate subjective initiative to protect ecosystems and more actively adopt GPBs. However, farmers are generally considered to have weak ecological awareness [32], with one of the reasons for rural ecological damage and environmental pollution being that farmers have low enthusiasm for ecological protection. As shown in Table 1, the indicator scores of farmers’ ecological awareness were generally higher than those of GPBs, indicating that although ecological policies had, to some extent, improved farmers’ ecological awareness, they were not sufficient to prompt them to take practical actions.

(2)

Positive subjective norms have been established around the protected areas, and some farmers actively responded to policies and adopted GPBs, seeking social recognition. Considering the characteristics of Chinese culture and rural areas, social identity theory may have stronger applicability in rural China. Farmers may be more susceptible to group influence and use “what we want to do” instead of “what I want to do” to guide their behavior; therefore, the long-term promotion of ecological policies seems more conducive to enhancing farmers’ adoption of GPBs from an intrinsic drive perspective.

(3)

At the same time, overemphasizing collectivism and mandatory constraints may have a negative impact on farmers’ GPBs, as shown by the fact that a stronger perception of the effect of ecological policies made it less likely that farmers would adopt GPBs. China is a country with a strong government and top-down management system, so farmers typically regard ecological protection as the government’s duty and believe that large-scale ecological crises can only be resolved through collective rather than individual efforts. That is, when farmers feel that policies have a beneficial effect, they may believe that their personal behavior has little impact and subsequently lose their enthusiasm for green production.

(4)

Farmers generally thought that they have weak perceived behavioral control related to their abilities and resources, so they consistently show resistance to change and a lack of motivation to adopt GPBs. To some extent, the weak perceived behavioral control of farmers is caused by the restrictions of ecological policies on local socioeconomic and farmers’ livelihood development. This seems to have negatively affected the psychological expectations of farmers, causing them to show a negative attitude toward ecological protection.

Table 4. Path analysis results.

Hypothesis	Impact Path			Estimated Value	Standard Error	C.R.	p Value	Validation Results
H1a	GPBs	←	EA	0.448	0.054	8.261	***	validated
H1b	GPBs	←	PE	−0.082	0.037	−2.220	0.026	reverse validated
H1c	GPBs	←	SN	0.330	0.050	6.667	***	validated
H1d	GPBs	←	PB	0.016	0.024	0.654	0.513	not verified
H2a	EA	←	PE	0.105	0.048	2.202	0.028	validated
H3a	EA	←	SN	0.493	0.057	8.660	***	validated
H3b	PE	←	SN	0.173	0.047	3.687	***	validated
H4a	EA	←	PB	0.104	0.032	3.294	***	validated
H4b	PE	←	PB	−0.027	0.032	−0.841	0.400	not verified

Note: *** indicates significance at a level of 1%.

Table 4. Path analysis results.

Hypothesis	Impact Path			Estimated Value	Standard Error	C.R.	p Value	Validation Results
H1a	GPBs	←	EA	0.448	0.054	8.261	***	validated
H1b	GPBs	←	PE	−0.082	0.037	−2.220	0.026	reverse validated
H1c	GPBs	←	SN	0.330	0.050	6.667	***	validated
H1d	GPBs	←	PB	0.016	0.024	0.654	0.513	not verified
H2a	EA	←	PE	0.105	0.048	2.202	0.028	validated
H3a	EA	←	SN	0.493	0.057	8.660	***	validated
H3b	PE	←	SN	0.173	0.047	3.687	***	validated
H4a	EA	←	PB	0.104	0.032	3.294	***	validated
H4b	PE	←	PB	−0.027	0.032	−0.841	0.400	not verified

Note: *** indicates significance at a level of 1%.

4.4. Indirect Effect Analysis

In the SEM, a direct effect represents the direct influence of one variable on another, whereas an indirect effect represents the influence of one variable on another through a third intermediary variable. If the direct effect is lower than the indirect effect, the intermediary variable plays an important role and requires attention [31]. The indirect effects of our analysis, shown in Table 5, indicated the following.

(1)

The direct effect of PE on GPBs was negative, but EA had a positive mesomeric effect (0.047), which ultimately alleviated the total negative effect. The implementation of a series of ecological policies has increased farmers’ perceived importance of and attention to ecological protection. However, they retained a lowered awareness of their own responsibilities, believing that their protection efforts are insignificant and that ecosystems should be protected by the government. As a result, farmers were not enthusiastic about adopting GPBs. Thus, it is important to recognize the dual characteristics of policy effects, striving to reduce their negative effects while increasing positive ones.

(2)

In addition to the direct effect that farmers’ SN had on GPBs, it also exerted a positive indirect effect (0.215) on GPBs through both PE and EA, ultimately maximizing its total effect. This shows that in a rural “acquaintance society”, information and communication from outside have a greater impact on farmers’ perception, attitude, and behavior, resulting in a significant correction effect on farmers portrayed as “rational economic men”. Farmers are always first constrained by mandatory and non-mandatory norms, after which they evaluate the policies that generate these constraints and thus influence other people’s ecological awareness affected by the spread of informal institutions. From the analysis results, it is evident that this impact was mainly positive and played a driving role in the adoption of GPBs. Policymakers should consider gradual, long-term changes to local customs to deeply embed the concept of sustainable development in people’s minds, thereby transforming external incentives into internal motivation.

(3)

Farmers’ PB had a positive indirect effect (0.048) on GPBs through PE and EA, which was even higher than its direct effect. Farmers’ perception and awareness represent the subjective judgment of their own conditions and capabilities. The higher the PB, the more resources farmers have, or the higher their capabilities to master them. This will facilitate their adoption of more environment-friendly values, improve their support for ecological policies, and thus stimulate them to make more altruistic behavioral decisions, that is, make a greater effort to adapt to GPBs. Although ecological policies do not directly subsidize farmers’ production activities, improving the local environmental infrastructure and enhancing their resources and learning abilities remain effective strategies for achieving successful ecosystem protection.

(4)

Farmers’ SN had a positive indirect effect (0.018) on EA through PE. Protection policies consistently manifest a positive subjective norm for farmers. The deeper their participation in and adaptability to ecological policies, the deeper their understanding of ecological policies, which can imperceptibly improve their environmental concerns and ecological awareness. This further enhances farmers’ support for these policies, forming a beneficial cycle. As a result, it is suggested that the participation of communities and farmers should be considered in the design of future ecological policies.

(5)

The direct effect of farmers’ PB on EA was positive; however, it also had a small negative indirect effect (−0.003) through PE, leading to a slight reduction in the total effect. Farmers with more forest land and other resource endowments also faced more pressure from ecological policies and were more susceptible to significant economic losses resulting from ecological protection. Thus, such farmers consistently exhibited a stronger negative perception of policies, which somewhat reduced their ecological awareness. Notably, this effect was relatively weak, primarily indicating the constraint of and risk associated with ecological policies affecting a specific, small group of farmers with significant resource endowments.

Table 5. Indirect effects as per SEM.

Hypothesis	Impact Path			Total Effect	Direct Effect	Indirect Effect
H1a	GPBs	←	EA	0.448	0.448	−
H1b	GPBs	←	PE	−0.035	−0.082	0.047
H1c	GPBs	←	SN	0.545	0.330	0.215
H1d	GPBs	←	PB	0.064	0.016	0.048
H2a	EA	←	PE	0.105	0.105	−
H3a	EA	←	SN	0.511	0.493	0.018
H3b	PE	←	SN	0.173	0.173	−
H4a	EA	←	PB	0.101	0.104	−0.003
H4b	PE	←	PB	−0.027	−0.027	−

Table 5. Indirect effects as per SEM.

Hypothesis	Impact Path			Total Effect	Direct Effect	Indirect Effect
H1a	GPBs	←	EA	0.448	0.448	−
H1b	GPBs	←	PE	−0.035	−0.082	0.047
H1c	GPBs	←	SN	0.545	0.330	0.215
H1d	GPBs	←	PB	0.064	0.016	0.048
H2a	EA	←	PE	0.105	0.105	−
H3a	EA	←	SN	0.511	0.493	0.018
H3b	PE	←	SN	0.173	0.173	−
H4a	EA	←	PB	0.101	0.104	−0.003
H4b	PE	←	PB	−0.027	−0.027	−

5. Conclusions and Policy Implications

The results presented here show that agroforestry farmers’ GPBs are affected by their ecological awareness, subjective norm pressures, and perceived behavioral control. Ecological policies have continuously improved the intensity and scope of ecological protection, which has profoundly affected farmers’ environmental efforts in contrasting ways. Specifically, farmers’ ecological awareness was improved by cultivating an atmosphere wherein rural societies focus on ecological protection, which in turn exerted positive regulatory pressures on farmers and changed their inherent customs and behavioral patterns. However, owing to the pressures of ecological policies on farmers’ resource endowment, their perceived behavioral control remained insufficient, which objectively hinders their adoption of GPBs. Therefore, we propose the following suggestions for optimizing future ecological policies.

(1)

It is crucial to enhance the ecological awareness of farmers by prioritizing environmental education over environmental publicity. Farmers’ ecological awareness remains at a basic level; they understand the importance of ecological protection, but there is no long-term understanding of the connection between ecological protection and economic and social development. This may be because ecological policies currently prioritize publicity over education. However, education would further inspire, while propaganda relies on indoctrination and is intentionally kept relatively simple to achieve higher dissemination rates. Specifically, we should help farmers understand the long-term relationship between environmental protection and economic development and encourage them to integrate their individual interests with the ecological welfare of society as a whole, which will enhance their subjective inclination to adopt GPBs. Furthermore, we should focus on educating farmers on their environmental responsibility, improving their participation in and adaptability to policies, and highlighting that their GPBs are effective for ecological protection.

(2)

The informal system plays a decisive role in the development and preservation of local order; therefore, the formal system should focus on restrictive measures, and ecological policies should utilize the informal system to implement ethical education and governance for farmers. We should also ensure to give more positive feedback to farmers who adopt GPBs, such as higher social status, recognition from the government, and respect from others. Building on the “herd effect”, this would encourage farmers to continually improve their protection efforts, thus affecting their GPBs.

(3)

Farmers’ perception of weak resources and capabilities to adopt GPBs strongly hindered the adoption of such behaviors; even if farmers’ ecological awareness had improved, their awareness deviated from their actual behaviors. According to the theory of “rational farmers” and the positive externality of green behavior, when the resources and abilities controlled by farmers are limited, they are more inclined to prioritize them over ecological protection to improve the family’s livelihood. Therefore, ecological policies should create more convenient conditions conducive to the adoption of GPBs. First, instead of emphasizing absolute ecological protection and disregarding the interests of farmers, it is crucial to encourage farmers to provide high-quality ecological products using rich ecological resources and adopt greener livelihoods. Second, compensation and security should be further increased while also enabling them to improve their management skills and learning abilities. Furthermore, we should improve the environmental infrastructure of rural communities, increase capital investment, form a standard system guarantee, and create convenient objective conditions enabling farmers to adopt GPBs.

(4)

Nevertheless, the fact that the adoption of GPBs is more time-consuming, laborious, expensive, and difficult to learn and master than traditional production methods must be considered. As noted previously, when facing the choice between economic development and ecological protection, most farmers’ inclination to choose the former should not be ignored. To alleviate this conflict, it is essential to help farmers abandon outdated traditional production concepts and actively explore mechanisms for realizing the value of ecological products, that is, goods produced utilizing excellent ecological endowments and green, pollution-free production methods. In addition, the government should guide the development of the ecological product market. Undoubtedly, when the benefits obtained from the production of ecological products exceed those of traditional products and the goals of ecological protection and economic development have been agreed upon, the adoption of GPBs will become an internal driving force for farmers.