Effect of ecoliteracy on farmers’ participation in pesticide packaging waste governance behavior in rural North China

Abstract

Farmers’ participation in pesticide packaging waste (PPW) governance is important for improving agricultural pollution and achieving sustainable agricultural development. By incorporating the theory of planned behavior, value-belief-norm theory, cognition and behavior theory etc., we construct a theoretical model comprising “ecoliteracy–farmers’ WTP in PPW governance–participation in PPW governance behavior.” This study investigates how ecoliteracy affects farmers’ participation in PPW governance and explores the mediating effect of farmers’ willingness to participate (WTP) in PPW governance. We use structural equation modeling to analyze data collected from a questionnaire survey including 1118 samples of Chinese farmers. The results show that (1) Ecoliteracy significantly affects farmers’ WTP in PPW governance. Ecological cognition, emotion, values, and knowledge and skills positively affect WTP in PPW governance, while ecological cognition and ecological knowledge and skills significantly affect participation in PPW governance behavior. (2) Farmers’ WTP in PPW governance mediates ecoliteracy and governance participation behavior. (3) Heterogeneity analysis reveals that different planting scales, different planting categories, and receiving/not receiving government project support have different effects on farmers’ participation in governance behavior. Farmers in the large-scale group are more likely to participate in governance than those in the medium- and small-scale groups, and farmers in the mixed grain and economic category are more likely to participate in governance than those in the economic and grain categories. Furthermore, farmers who receive government support are more likely to participate in governance than those who do not. Our results can serve as a policy making reference for promoting PPW governance in various regions.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-024-73858-y.

Introduction

Pesticides are important for improving agricultural productivity and ensuring global food security^1,2. However, with the massive use of pesticides in rural areas, pollution from pesticide packaging waste (PPW) has become a major issue³. China, a large agricultural country, produces up to 10 billion pesticide packages annually, more than three billion of which are casually discarded, comprising approximately 100,000 tons of waste⁴. Pesticide packaging is often made of glass or plastic and is not readily degradable⁵, resulting in the deterioration of the agricultural production environment⁶. Moreover, PPW often contains 2–5% pesticide residue^7,8, which can affect the safety of the food supply chain. It is urgent, then, to improve PPW management to support green agricultural development^9,10.As important participants in PPW treatment, farmers are the main link for improving its efficiency¹¹. China has contractually cooperated with farmers through project-based policies and public service outsourcing to standardize the performance of farmers’ duties^12,13. Although regulations have produced some achievements in PPW treatment, farmers’ enthusiasm for PPW participation is generally low¹⁴, it is difficult to stimulate the initiative of farmer; and the recovery rate is less than 15%¹⁵, sustained participation in the management of pesticide packaging waste is not enough incentive. The reason for this is that PPW treatment is a public good for which farmers typically pay no cost, and free-riding behavior subsequently emerges¹⁶, which makes it difficult for individuals and collectives to achieve effective cost-sharing, resulting in collective action dilemmas¹⁷. In addition, approximately 63% of cultivated land in China is operated by approximately 200 million small farmers¹⁸, and small-scale, decentralized, single-household operation increases the difficulty of PPW treatment¹⁹. Thus, it is important to determine how to improve farmers’ initiative and participation in PPW treatment.

In 2024, the Ministry of Agriculture and Rural Development issued the “Opinions on Strengthening the Work of Highly Skilled Agricultural and Rural Talents in the New Era”, which proposes to improve the overall quality of the highly skilled agricultural and rural talent team. In the context of building a rural ecological civilization and strengthening the development of a highly skilled rural construction workforce, there is an urgent need to improve the quality of farmers’ qualities, especially in order to comprehensively promote the implementation of the governance of pesticide packaging waste, and to develop the capacity of farmers for self-development. According to the theory of farmers’ behaviour and the theory of motivation, farmers’ participation in environmental management behaviors is not entirely exogenous, and there is also a diverse range of protective behaviors for farmland farming systems. Farmers’ behaviors are driven by intrinsic motivation, and ecoliteracy can enhance intrinsic motivation for environmental behaviors²⁰. Numerous scholars have also begun to pay attention to the impact of ecological literacy on farmers’ behavior in participating in pesticide packaging waste management, the related research has mainly focused on, first, ecoliteracy’s effect on individual environmental protection behaviors. Using ecoliteracy models, researchers have formed research paradigms based on factors such as ecological cognition, emotion, values, and knowledge and skills^21–23. These can serve as self-restraints for individuals and promote their environmental protection behaviors²⁴. For example, farmers’ ecological cognition and skills positively influence their agricultural environmental protection behaviors²⁵. High awareness of responsibility can convert farmers’ willingness to participate (WTP) in environmental improvement into actual behavior²⁶. Positive environmental attitudes can, moreover, promote waste disposal behavior by individual households²⁷. Second, studies have investigated the influence of psychological factors on farmers’ PPW governance behaviors. Behavioral economics theory suggests that decision-making behavior is characterized by uncertainty; in this regard, psychological factors can be used to scientifically analyze decision-making behavior^28,29. Studies have explored the various influence paths using the theory of planned behavior and its extensions. For example, using perceived value theory, Li³⁰ found that perceived benefit was the most important factor influencing farmers’ green waste disposal behaviors. Meanwhile, Lina³¹ used attitude, subjective norms, perceived behavioral control, and other internal psychological mechanisms to identify the behavioral logic of PPW governance among fruit farmers. Using an environmental social psychology approach, Cai³² found that supportive attitudes positively affected PPW recycling among rice farmers in Guangdong Province. Wang³³ found that increases in farmers’ ecological cognition made them more inclined to adopt PPW recycling. Others, meanwhile, have focused on the effect of endowment heterogeneity on farmers’ behavioral decisions based on factor endowment theory. Whether farmers implement PPW governance depends on household resource endowment^34,35, education³⁶, and income³⁷, and differences in endowment can lead to different behavioral responses.

While existing studies can provide a reference for the present study, there is room for supplementation and improvement. First, the decision-making behind farmers’ PPW behavior is a complicated process.The existing research analyzes the impact of farmers’ participation in PPW recycling behavior, which is one of them, and does not embed farmers’ behavior in the whole chain of “organization-collection-disposal” of PPW governance. At the micro level of farmers, the treatment of pesticide packaging waste can be classified as recycling behavior, centralized disposal behavior, community supervision behavior. The conclusions drawn from this can not accurately explain the key factors affecting farmers’ participation in treatment behavior, and can not fully reflect the current situation of farmers’ PPW behavior. Second, most studies analyze ecoliteracy and farmers’ environmental governance behavior as independent research objects and do not incorporate ecoliteracy and PPW governance behavior into a unified framework. ecoliteracy is highly concerned in the field of environmental education. Some research has evaluated the level of ecoliteracy and explored its impact on individual environmental behavior, but little was known about the level of farmers’ ecoliteracy and its influence on farmers’ PPW behavior. At present, scholars have not sufficiently discussed the formation mechanism of this behavior. There is no clear explanation about the impact patterns, formation processes and interactions of each variable with environmental behavior. Third, studies of PPW governance tend to treat farmers as a homogeneous group, ignoring heterogeneity in their PPW governance behavior. In light of the above, this study applies structural equation modeling (SEM) to data from 1,118 farmers in Hebei Province, China, to empirically investigate ecoliteracy’s effect on PPW governance participation. We further reveal heterogeneity in farm household PPW governance participation, aiming to provide empirical evidence and a decision-making reference for promoting PPW governance and the green transformation of agriculture.

This study’s contributions are as follows. Firstly, in the context of the rural cultural revitalization era, it analyzes farmers’ PPW from the perspective of ecoliteracy, offering a novel viewpoint to promote farmers’ PPW. And, it enriches the measurement of farmers’ PPW by assessing it in three phases: organization→recycling behavior, collection→ centralized disposal behavior, disposal→community supervision behavior, providing a comprehensive reflection of farmers’ PPW status, considering the whole chain of PPW management. Secondly, by incorporating the Theory of Planned Behavior, Value-belief-norm theory, Cognition and behavior Theory etc., we construct a theoretical model comprising “ecoliteracy–farmers’ WTP in PPW governance–participation in PPW governance behavior.” We establish an ecoliteracy indicator system in four dimensions: ecological cognition, ecological emotion, ecological values, and ecological knowledge and skills. Into the analysis of the impact mechanism of ecoliteracy on farmers’ PPW, it clarifies the relationships among ecoliteracy, farmers’ willingness to participate in PPW governance, and farmers’ PPW through SEM effect tests, revealing new pathways to promote farmers’ engagement in PPW. Thirdly, through the SEM effect test and heterogeneity analysis, we clarify the impact mechanism of ecoliteracy on farmers’ PPW behavior. We explore the effect of farmers’ heterogeneity on their PPW governance participation. Farmers are grouped according to cultivation category, cultivation scale, and whether they receive government support; Multi-group SEM is used to clarify the differences between the influence pathways. Therefore, the research questions of this paper are focused on farmers, aiming to investigate whether ecoliteracy, as a critical factor, can influence farmers’ PEB, how this influence occurs, and the differences in its impact on various PPW among different farmers. The goal is to better promote the participation of PPW by farmers and to reduce the environmental damage caused by agricultural production. Furthermore, it explores the differential impact of farmers’ ecoliteracy on different types of PPW, providing insights for formulating targeted green production guidance policies.

In the present study, we present the study’s theoretical framework and hypotheses in Sect. Theoretical framework and hypothesis. Section Materials and methods describes the research method, and Sect. Results presents the results. Section Discussion discusses the findings, and Sect. Conclusions concludes this paper.

Theoretical framework and hypothesis

Ecoliteracy was introduced by Orr³⁸, who suggested that it includes the basic competencies of reading and learning, knowledge use, and systematic thinking, combining the “natural landscape” and the “mind landscape” and emphasizing coordinated interaction in the “human–social–ecological system³⁹”. We define ecoliteracy as the social quality of farmers that enables them to acquire environmental protection knowledge through social learning, environmental problem-solving skills, ecological values, agricultural production practice, informed decision-making, and identifying, analyzing, and solving problems. PPW governance participation is the ultimate expression of farmers’ ecoliteracy and the concrete expression of ecological cognition, emotions, values, and knowledge and skills in PPW governance. The discussion on the connotation of ecoliteracy is mainly shown in Table 1.

Table 1.

Conceptual framework for Ecoliteracy.

Representative	Connotation
David W. Orr (1992)	Ecoliteracy: education and the transition to a post modern world: Ecoliteracy includes the basic competences of reading and learning, knowledge use, and systems thinking, and combines ‘natural landscapes’ and ‘mental landscapes’.
Capra (1997)	The web of life : a new scientific understanding of living systems: Ecoliteracy is a necessary quality for members of society, and is important for the sustainable development of society and the construction of life networks; people with ecoliteracy need to have a wealth of knowledge and a caring attitude towards the environment, as well as the ability to take action in practice.
Stable(1998)	There are three levels of ecoliteracy: functional ecoliteracy (the ability to understand ecological regions, ecological ‘facts’ and landscapes), cultural ecoliteracy (the ability to understand the cultural significance of nature’s imagery and to capture human dimensions and roles in landscapes), and critical ecoliteracy (the ability to actively explore the significance of the environment for oneself and others and to develop an understanding of how to improve the environment through action). Critical ecoliteracy (actively exploring the meaning of the environment for oneself and others and developing an understanding of how to improve understanding of the environment through action).
Rosalyn (1999)	Ecoliteracy means that the subject has knowledge of ecosystems, understands the interplay between human beings and nature, is able to practice eco-friendly behaviors and has the ability to analyse and solve problems.
Bruyere (2008)	Ecoliteracy consists of three components: ecological knowledge, attitudes and behaviors
China Project Team (2010)	Research on the Indicator System for Assessing the Environmental Quality of the Chinese Public: Ecoliteracy include: environmental knowledge, environmental values, environmental attitudes and environmental behaviors.
Davidson (2011)	There are three main levels of ecoliteracy: environmental attitudes, environmental values, and understanding of environmental issues.
McBride (2013)	Ecoliteracy includes six sections on life cycles and biological networks, ecosystem services, negative human impacts on ecosystems, critical thinking and application, the nature of ecological sciences and bio-geography, with a strong emphasis on ecological knowledge, and interactions between humans and ecosystems.

Influence of ecological cognition on farmers’ participation in PPW governance behavior

Ecological cognition is essential for farmers to generate PPW governance behavior preferences⁴⁰. In terms of subjective logic, ecological cognition is the basis of farmers’ participation in governance behavior. According to cognitive behavioral theory, farmers’ behaviors are related to cognition, which affects behavioral decisions at the conscious level. Cognition has a direct impact on an individual’s behaviour and a change in cognition ultimately triggers a change in behaviour. Ecological cognition is the process of understanding, processing and restructuring of environmental information, including the subject’s actions on the perception of the environmental impact of the perception of the importance of protecting the rural environment, the importance of the judgement of whether it can meet the needs of the development of the outside world through the thinking or sensory processing of information. False cognition is the main source of bad behavior, while correct cognition is the main source of good behavior. Based on this theory, PPW cognition may have positive influence on farmers’ PPW behavior. Indeed, previous studies have widely confirmed that cognition has an important effect on farmers’ environmental protection behaviors. Wang et al. found that farmers’ ecological cognition can significantly promote their PPW management behaviors; the higher their level of cognition, the greater the likelihood of adopting environmental management behavior⁴¹. Zhao⁴² showed that farmers’ cognition of the recycling and management of plastic waste improved their participation. Garud⁴³ argued that in PPW governance, actors’ subjective initiative is important for breaking through path dependence; rational environmental perceptions can motivate farmers to actively participate in packaging waste management and pesticide waste collection. Meanwhile, ambiguous environmental perceptions might cause them to make incorrect decisions and generate random packaging waste discarding behavior. Thus, we propose the following hypotheses:

1. Ecological cognition positively affects farmers’ WTP in PPW governance.

2. Ecological cognition positively affect farmers’ participation in PPW governance behavior.

Influence of ecological sentiment on farmers’ participation in PPW governance behavior

The implementation of PPW governance behavior by farmers as “rational economic beings” is influenced by not only ecological cognition but also attitudinal factors. Theory of ‘knowledge-emotion-intention-action’ holds that human behaviour depends on the three basic elements of knowledge, emotion and intention, of which ‘emotion’ is sentiment and mood, which promotes ecological conservation behaviors by influencing the motivation, direction and continuity of individuals. Ecological emotion is a collection of attitudes, beliefs, and behavioral intentions related to environmental matters⁴⁴. Watson and Tellegen⁴⁵ classified ecological emotions into positive and negative emotions. Positive emotions are reflected in the positive emotions generated by farmers’ adoption of environmental protection behaviors; their sense of responsibility can enhance their ecological protection cognition, which will stimulate their sense of pride and self-identity, prompting them to adjust their behaviors⁴⁶. Negative emotions are manifested in a sense of guilt, which can affect ecological protection behavior⁴⁷ this, in turn, will reduce the pressure brought by negative emotions. Farmers will improve their behavior. Other studies have analyzed the relationship between cognition and emotion using the “knowing–believing–acting” model, suggesting that cognition is the basis of emotion, where there is a progressive relationship from cognition to emotion⁴⁸. Here, ecological emotion will stimulate farmers’ motivation to participate in environmental management. Sometimes, ecological emotion is more effective than improving farmers’ cognition⁴⁹. Thus, we propose the following hypotheses:

• H2aEcological emotion positively affects farmers’ WTP in PPW governance.
• H2bEcological emotion positively affects farmers’ participation in PPW governance behavior.

Influence of ecological values on farmers’ participation in PPW governance behavior

Ecological values refer to individuals’ fundamental views about environmental issues. Farmers’ environmental behaviors are significantly influenced by values, and there is a causal relationship between values and pro-environmental behaviors⁵⁰. Stern’s value–belief–normative theory suggests that altruistic values and ecological values act on an individual’s cognitive system, which can better regulate an individual’s environmental behavior⁵¹. Behavioral science suggests that values underlie the formation of attitudes and that they are more stable and enduring predictors of behaviour than attitudes. Farmers’ ecological values directly affect their participation behaviors; those with stronger ecological values are more willing to adopt green consumption and pro-environmental behaviors⁵². Altruistic values, moreover, are essential for promoting trust among farmers and the spillover of circular agriculture technologies⁵³. Meanwhile, values moderate the effect of environmental attitudes on environmental behaviors. Farmers are also regulated and constrained by ecological values when participating in PPW management, and differences in individual behavioral decisions are affected by differences in values. Accordingly, we propose the following hypotheses:

• H3aEcological values positively affect farmers’ WTP in PPW governance.
• H3bEcological values positively affect farmers’ participation in PPW governance behavior.

Influence of ecological knowledge and skills on farmers’ participation in PPW governance behavior

Schultz considers skills to be particularly important among all human resources and, in book ‘Human Capital Investment and Urban Competitiveness’, argues that differences in the skills, experience and proficiency of workers manifest themselves in different productive capacities and behavioral decisions. Ecological knowledge and skills refer to the ability to understand and master the content and requirements of environmental protection and to identify and solve environmental problems⁵⁴. Individuals can only develop ecological awareness after acquiring certain knowledge and skills. Farmers with rich knowledge tend to hold positive ecological values, are more cognizant of the adverse effects of environmental pollution, and ultimately activate personal norms⁵⁵. Ecological knowledge can also help farmers obtain accurate information and resources, participate in training to master environmental management skills, change their environmental capacity, and improve their behavioral decision-making ability⁵⁶. In economics, the knowledge and skills acquired by individuals are their most important resources, and skill acquisition bias affects behavioral decision-making and behavioral capacity⁵⁷. Studies have confirmed that the degree of the implementation of green production behaviors by farmers participating in ago-technology training is higher than that by those who do not participate in such training⁵⁸. Farmers’ acquisition of appropriate skills to deal with ecological problems on farmland will stimulate the emergence of a willingness to conserve ecology, which in turn will guide their ecological conservation behaviour on farmland. Farmers’ search for information and ecological knowledge to increase access to resources has the potential to correct biased information on action strategies in order to motivate them to act more responsibly. Accordingly, the following hypotheses are proposed:

• H4aEcological knowledge and skills positively affect farmers’ WTP in PPW governance.
• H4bEcological knowledge and skills positively affect farmers’ participation in PPW governance behavior.

Mediating effects of farmers’ WTP in PPW governance behavior

Based on the theory of planned behavior, value norm theory, and cognitive behavioral theory, willingness is the most direct influence on behaviour, studies have shown that ecological knowledge and skills enhance the relationship between willingness to dispose of recycled waste and behavior⁵⁹. Willingness has a transmission function between ecological literacy and farmers’ behavior. Individuals guide themselves to implement ecological behavior on the basis of establishing behavioral willingness, that is, willingness is the preceding paragraph of behavior. Farmers’ behavioral intentions drive ecological conservation behavior, and those with behavioral intentions are more likely to make behavioral decisions. Ecoliteracy reduces the likelihood of behavioral deviation. Subjective norms and attitudes, green perceptions, and perceived behavioral control promote rural residents’ green technology adoption and willingness to classify domestic waste^60,61. Willingness also has a transmission function between ecoliteracy and farmers’ behaviors. Digital literacy can significantly enhance farmers’ WTP in the governance of domestic waste classification⁶². The relationship between faucal sludge treatment technology and training plays a positive role in promoting farmers’ WTP⁶³. Farmers’ ecoliteracy significantly affects their pro-environmental willingness and behavior. Ecoliteracy is shaped by agricultural production practices and interactive learning exchanges, which stimulate their willingness to protect farmland and implement governance behaviors⁶⁴. We therefore propose the following hypothesis:

• H5Farmers’ WTP in PPW governance mediates the relationship between ecoliteracy and farmers’ participation in PPW governance behavior.
• H5aFarmers’ WTP in PPW governance mediates the relationship between ecological cognition and farmers’ participation in PPW governance behavior.
• H5bFarmers’ WTP in PPW governance mediates the relationship between ecological emotion and farmers’ participation in PPW governance behavior.
• H5cFarmers’ WTP in PPW governance mediates the relationship between ecological values and farmers’ participation in PPW governance behavior.
• H5dFarmers’ WTP in PPW governance mediates the relationship between ecological knowledge and skills and farmers’ participation in PPW governance behavior.

In summary, we construct a theoretical framework made up of ecoliteracy–farmers’ WTP in PPW governance–farmers’ participation in PPW governance behavior. Figure 1 shows the theoretical framework. This study explores the direct effect of ecoliteracy on farmers’ participation in PPW governance behavior and their WTP in PPW governance; meanwhile, the mediating effect of WTP in PPW governance on ecoliteracy’s effect on farmers’ participation in PPW governance is also investigated. The purpose of this research is to explore the influence mechanism of each hypothesis path.

Fig. 1.

Theoretical framework.

Materials and methods

Data collection

Our data come from questionnaire surveys conducted from June to July 2021 and June to September 2022 in Cangzhou, Baoding, Tangshan, Zhangjiakou, and Handan, which are the main agricultural production areas of Hebei Province. The study area was selected with the following considerations: First, Hebei Province, as a major agricultural province, is an important supply base for agricultural and sideline products for Beijing and Tianjin. According to the China Rural Statistical Yearbook 2022, the total sown area of crops in the province in 2021 was 8,097,200 hectares, which is among the highest in the country. Second, the study area is mainly a provincial pilot area for PPW recycling and treatment, which is well represented in the study of farmers’ participation in PPW governance behavior. After the establishment of national-level PPW recycling and treatment pilot counties in 2019, in March 2020, Hebei Province also adopted a pilot project for PPW recycling and treatment in 14 large grain (oil) producing counties and key counties of the vegetable industry, such as Qingxian County in Cangzhou, Anguo in Baoding, and Zhangbei County in Zhangjiakou. Moreover, to examine the heterogeneity of farmers’ participation in governance behaviors under environmental regulations, we also select non-provincial pilot regions such as Luannan County in Tangshan and Chicheng County in Zhangjiakou. Third, the study area is a major agricultural production area, and the demand for PPW governance continues to grow. The study area covers the main production areas of cash crops or food crops, such as melon and vegetable in Qingxian County, the facility vegetable industry in Luannan County, staggered-season vegetable production in Zhangbei County, wheat-corn cultivation in Yongnian District, and traditional Chinese medicinal materials in Anguo. Figure 2 shows the regional distribution of the research.

Fig. 2.

Study area distribution. (Note: The review number of the Chinese map used this time is GS (2019)1822, and the base map has not been modified.)

Our research process combines typical sampling and random sampling. First, based on the construction of PPW recycling and disposal pilot counties, we take five cities and 10 counties (districts) as the study areas: Qing County, Cangzhou; Anguo, Gaobeidian, and Zhuozhou in Baoding; Qianan and Luannan County, Tangshan; Zhangbei County and Chicheng County, Zhangjiakou; and Yongnian District and Jize County in Handan. Second, we randomly select two to three administrative villages from each township and randomly choose 10–15 farmers from each for the survey. The survey sample covers 87 villages. The survey includes 1,190 samples of farm households, excluding those missing key information and abnormal data samples. We finally obtain 1,118 valid questionnaires, for a sample validity rate of 93.95%.

The sample characteristics mainly include the three dimensions of individual characteristics, production characteristics, and household characteristics. Individual characteristics include gender, age, and education level(Table S1). Among the 1,118 farm households interviewed, male respondents are predominant, and the main labor force in agricultural production is male. There are 740 male respondents, accounting for 66.19%, and 378 female respondents, accounting for 33.81%. The age of the interviewees is mainly concentrated around 56–65 years old, with a total of 494 people, accounting for 44.19%; population aging in agricultural cultivation is serious. Overall education level is low, with the mean value located at the junior high level; interviewed farmers with an education level of junior high school and below total 881 people, accounting for 78.80%. Production characteristics include planting years, planting scale, and number of plots; 51.70% of the farmers have been planting for more than 35 years, and 39.35% have cultivated an area of less than three acres of land. Land is highly fragmented, with an average of three plots. The number of farmers owning one to three plots of land is 792, accounting for 70.84% of the total number of interviewed farmers. Household characteristics are mainly the proportion of agricultural income to total household income, with 90.43% of agricultural income accounting for more than 70% of household income. The most important source of rural household business income is still mainly farming.

Measurement of variables

To ensure the reliability and validity of the questionnaire, the selection of variables is based on the concept of ecoliteracy, and mainly derived from the existing literature, latent variables quantified in this study have been slightly modified according to the research topic. In summary, we take “ecological cognition,” “ecological emotion,” “ecological values,” and “ecological knowledge and skills” as endogenous latent variables. “Ecological cognition” is measured based on farmers’ understanding and perception of PPW governance, including four dimensions: time and energy cognition, management effect cognition, pollution prevention cognition, and health protection cognition. The Ecological cognition was mainly developed from the work of Xue et al. (2022)⁶⁵ and Ren et al.⁶⁶. “Ecological emotion” measurement variable is divided into positive ecological emotion and negative ecological emotion, with a total of three items for observation. Positive ecological emotion is observed based on the sense of responsibility and pleasure related to PPW recycling; negative emotion is observed based on the sense of ecological guilt, aiming to judge farmers’ awareness of the effects of not recycling PPW. Negative emotions are observed through ecological guilt to determine farmers’ psychological guilt for not carrying out PPW recycling behavior. The Ecological cognition was mainly developed from the work of Junaedi⁶⁷ and Xue et al.⁶⁸. “Ecological values” is measured based on harmonious coexistence between human beings and nature, PPW can be recycled by farmers; Green agricultural development, Concern for the rural environment. The Ecological values was mainly developed from the work of Lamarque et al.⁷⁰, Maurer and Bogner⁷¹ and Yu et al. ⁶⁹ “Ecological knowledge and skills” measured based on the understanding of the governance process, policies and regulations, and problem-solving ability. The Ecological knowledge and skills were mainly developed from the work of Tuncer et al.⁷² and Ramdas and Mohamed⁷³.

Based on the theory of planned behavior, “willingness to participate in governance” and “participation in governance behavior” are selected as exogenous latent variables. Among them, “willingness to participate in governance” is reflected by farmers’ willingness to recycle, use centralized disposal, and conduct supervision and management. “Participation in governance behavior” is reflected by farmers’ recycling, centralized disposal, and supervision and management behavior. In Table 2, the measurement questions are all based on five-point Likert scales (1 = strongly disagree, 2 = comparatively disagree, 3 = generally disagree, 4 = comparatively agree, and 5 = strongly agree).

Table 2.

Definition, item, code, and descriptive statistics of variables.

Latent variables	Measurement item	Mean	Std.	Code
Ecological cognition (EC)	Have enough time and energy to participate in PPW treatment	3.71	1.114	EC1
	Participating in recycling can promote PPW treatment	3.83	1.046	EC2
	Be responsible for preventing and controlling PPW pollution	3.57	1.108	EC3
	Be responsible for maintaining the health of the public	3.75	1.129	EC4
Ecological emotion (EE)	Have the responsibility and obligation to recycle PPW	4.13	0.969	EE1
	Participating in the recycling of PPW can bring pleasure	4.20	0.927	EE2
	There will be guilt if you do not recycle PPW	3.97	1.027	EE3
Ecological values (EV)	Harmonious coexistence between human beings and nature, PPW can be recycled by farmers	3.79	1.078	EV1
	Agricultural development should pay attention to ecological protection	3.86	1.065	EV2
	Pay attention to collective interests and environmental interests	3.60	1.106	EV3
Ecological knowledge and skills (ES)	Understand the process of PPW treatment	4.02	1.034	ES1
	Clarify the policies, laws, and regulations related to PPW	4.09	0.989	ES2
	Can find out the problems in PPW treatment in time	3.83	1.071	ES3
Willingness to participate in PPW governance (WTP)	Willing to participate in the recycling process of PPW	3.73	1.091	WTP1
	Willing to participate in the centralized disposal process of PPW	3.88	1.052	WTP2
	Willing to participate in the community supervision process of PPW	3.60	1.106	WTP3
Participation in PPW governance behavior (TPB)	Reasonable collection of PPW	4.02	0.996	TPB1
	Correct centralized disposal of PPW	3.87	1.048	TPB2
	Implement community supervision of PPW	3.81	1.083	TPB3

Methods

SEM is suitable for revealing complex causal relationships among multiple variables. Based on the covariance matrix of the variables, it can obtain the indirect, direct, and total effects of the independent variables on the dependent variable and can deal with measurement and analysis at the same time⁷⁴. This study aims to verify the direct effect of ecoliteracy on farmers’ participation in PPW governance behavior and whether ecoliteracy affects management behavior through the WTP in management; thus, structural equations are used for verification. SEM consists of two parts: the structural model consisting of latent variables and their relationships, and the measurement model consisting of latent and observed variables⁷⁵. The structural model part reflects the intrinsic relationship between the latent variables, and the measurement model part reflects the relationship between the observed variables and the latent variables. After analyzing the relationships among variables, the latent variables selected in this study are “ecological cognition,” “ecological emotion,” “ecological values,” “ecological knowledge and skills,” “farmers’ participation in PPW governance behavior,” and “willingness to participate in PPW governance.” These are then reflected by several observable variables. By establishing the path relationship between latent variables, we can test whether the sample data and the theory fit together and finally verify the research hypotheses. SEM method is applicable to the study of farmers’ behavior among multivariate variables, and scholars have successfully verified the research hypotheses using this method, as shown in Table 3.

Table 3.

Selected literature using SEM to Validate Farmer Behavior.

Research topics	Data sources	Authors and Journals
Farmers’ Low-Carbon Production Behavior	Actual process of distributing the questionnaires, face-to-face interviews, take the county as the unit for investigation	Wang et al(2024). Original Research
Environmental awareness affects adoption of greener production systems	275 winegrowers from the Conegliano Valdobbiadene Prosecco Protected Designation of Origin area	Maria et al(2024). Journal of Environmental Planning and Management
Conceptualization of Farmers’ Water Conservation Intention and Behavior	36,183 Iranian farmers, 380 of whom were interviewed in-person in the form a cross-sectional survey research design	Naser et al(2023). Water
Green Production Behaviors of Farmers Living in or near Protected Areas	One-to-one structured questionnaire surveys, 975 responses from 79 villages were collected, with 10–30 households	Shou et al(2023). Forests
Farmers’ Watershed Ecological Protection Behaviour	Field household survey, select two cities and four counties (districts), total of 750 questionnaires were distributed	Zhang et al(2023). Int. J. Environ. Res. Public Health
Factors affecting the willingness of agricultural green production from the perspective of farmers’ perceptions	Field survey, 12 townships (towns) and 24 administrative villages were selected from 6 sample counties	Li et al(2020). Science of The Total Environment
Green consumer behaviour	Self-administered questionnaire, uestionnaire available electronically using Survey Monkey before collecting the data	Armindaet al(2019). Journal of Cleaner Production
Adoption of Green Fertiliser Technology	Using a stratified sampling technique, 600 responses were taken from farmers in three different major paddy plantation	Adana et al(2018). Land Use Policy
Ecological conservation behavior of farmers in payment for ecosystem service programs	Five to seven villages were randomly selected in each county on the basis of total population, and respondents from 25 to 35 households per village were randomly selected from the list provided by the village head	Deng et al(2016). Sci Total Environ

Based on the above theoretical analyses, and considering the basic principles of structural modeling, we construct a structural model of the path relationship of the important latent variables in the hypotheses. The factors that directly affect farmers’ participation in PPW governance behavior include ecological cognition, ecological emotion, ecological values, ecological knowledge and skills, and WTP in governance. Ecoliteracy also indirectly affects farmers’ participation in governance behavior through their WTP in PPW governance, which shows a chain influence path. Referring to Wolf⁷⁶ and other studies, we add the observed variables of each latent variable and combine the structural part with the measurement part; the SEM of ecoliteracy’s effect on farmers’ participation in PPW governance behavior is then obtained. In Fig.S1, ξ denotes the exogenous latent variable, η denotes the endogenous latent variable, γ denotes the influence coefficient of the exogenous latent variable on the endogenous latent variable, and β is the influence coefficient among the endogenous latent variables. ζ denotes the structural residual term of the endogenous latent variable; EC, EE, EV, and ES denote the vectors of the exogenous observed variables; WTP and TPB denote the vectors of the endogenous observed variables; δ and ε denote the measurement error terms of the exogenous observed variable and the endogenous observed variable, respectively; and Φ denotes the exogenous correlation of the latent variables.

In the SEM of ecoliteracy’s effect on farmers’ participation in PPW governance behavior, the basic model can include a structural part and a measurement part. The structural part of the model can be written as.

1

The model of the measurement part can be written as.

2

Results

Pretest of the model

To determine whether the selection of variables is feasible, we use Stata 17.0 and AMOS 22.0 to perform scale reliability and validity tests, differentiation validity tests, and model goodness-of-fit tests.

Reliability and validity test

Following Nunnally⁷⁷, we test the questionnaire for reliability and validity using validated factor analysis. The test indices mainly include factor loading, Cronbach’s α, component reliability (CR), and convergent validity (AVE). Table S2 shows that the factor loadings of “ecological cognition,” “ecological emotion,” “ecological values,” “ecological knowledge and skills,” “WTP in governance,” and “participation in governance behavior” are above 0.7, indicating that all six dimensions can be effectively responded to by the measurement indicators. The Cronbach’s α coefficients of each dimension are above the standard value of 0.7, which indicates that the internal consistency of each measurement topic within a single dimension is good. The CR value is the combination of the reliability of all questions. The CR values of the exogenous latent variables “ecological cognition,” “ecological emotion,” “ecological values,” and “ecological knowledge and skills” are all above the standard value of 0.7 (0.844, 0.795, 0.806, and 0.798, respectively). The CR values of “WTP in governance” and “participation in governance behaviors” are 0.794 and 0.797, respectively, which are above the 0.7 threshold, indicating high internal consistency. The AVE is used to calculate the average explanatory ability of latent variables for the measurement indicators. The AVE of each dimension is greater than 0.5, which indicates that the model has good convergent validity.

Discriminant validity test

According to Tihomir⁷⁸, interconceptual discriminant validity is considered when the squared value of the AVE construct is greater than the correlation coefficient between the remaining constructs. As shown in Table S3, the open square values of the AVEs of the constructs of “ecological cognition (EC),” “ecological emotion (EE),” “ecological values (EV),” “ecological knowledge and skills (ES),” “farmers’ participation in PPW governance behavior (TPB),” and “willingness to participate in PPW governance (WTP)” are greater than the correlation coefficients of the other constructs; thus, the model has good discriminant validity.

Model goodness-of-fit test

Model fit is used to test whether the hypothetical theoretical model and the real data fit can match each other. Here, we select the parsimony fit index ; dissimilarity indices SRMR and RMSEA; and similarity indices TLI, CFI, and IFI. We use AMOS 24.0 to evaluate goodness of fit. is very sensitive to sample size, resulting in the need to combine the degrees of freedom to form the newly obtained indicator for analytical judgment in the analysis of large sample sizes greater than 200. According to the evaluation criteria of the overall fitness evaluation indices of SEMs given by Jackson⁷⁹, a value of 1–3 indicates that the model fit has better similarity indices. The similarity indices TLI, CFI, and IFI need to be directly proportional to the fit of the model, and the fitness criterion should be higher than 0.9. The values of the indicators SRMR and RMSEA need to be inversely proportional to the model fit; less than 0.05 is the standard, and 0.08 is acceptable. Table S4 shows the test results for the overall fitness of the theoretical model. Referring to the reference values of the above indicators, is 2.241, which is between 1 and 3; SRMR and RMSEA are 0.025 and 0.033, respectively, which are less than 0.080; and GFI, ACFI, IFI, CFI, and TLI are all over 0.9, indicating that the framework and sample data fit are inversely related to the model fit, and less than 0.05 is acceptable. This indicates that our research framework is appropriate for the sample data and can be further used to test the hypotheses.

Structure results

Based on our research assumptions, we construct an SEM diagram of ecoliteracy, farmers’ WTP in PPW treatment, and their behavior of participating in PPW treatment using AMOS 22.0. Figure 3 shows the standardized path results between the different dimensions of ecoliteracy and farmers’ PPW WTP and behaviors.

Fig. 3.

Structural equation model of farmers’ participation in PPW governance behavior.

In the path relationship of ecoliteracy’s effect on farmers’ participation in PPW governance behavior, ecological cognition, ecological emotions, ecological values, and ecological knowledge and skills have a positive and direct effect on farmers’ WTP in PPW governance. Table 4 shows that the standardized path coefficients are 0.306, 0.219, 0.121, and 0.350, and the p-values are all under 0.05, indicating that the hypotheses are all reasonable; thus, H1a, H2a, H3a, and H4a are verified. The standardized path coefficients corresponding to the three dimensions of ecological cognition, ecological knowledge and skills, and farmers’ WTP in governance are 0.128, 0.190, and 0.387, respectively, and are all significant at the statistical level of 0.1%, which indicates that they significantly and positively influence farmers’ participation in PPW governance behavior. Hypotheses H1b, H4b, and H5c are verified. Meanwhile, ecological emotions and ecological values do not pass the significance test, with standardized path coefficients of 0.071 and − 0.002, respectively, and H2b and H3b thus do not pass the test.

Table 4.

Parameter estimation results for the structural equation model effect paths.

Relationship in latent variables	Unstd.	S.E.	Z-value	Sig.	Std.	Significance
H1a: EC→WTP	0.292	0.034	8.549	***	0.306	Established
H1b: EC→TBP	0.128	0.040	3.229	0.001	0.128	Established
H2a: EE→WTP	0.216	0.035	6.240	***	0.219	Established
H2b: EE→TBP	0.073	0.040	1.849	0.064	0.071	Failed
H3a: EV→WTP	0.115	0.031	3.775	***	0.121	Established
H3b: EV→TBP	−0.002	0.034	−0.046	0.963	−0.002	Failed
H4a: ES→WTP	0.324	0.036	9.022	***	0.350	Established
H4b: ES→TBP	0.185	0.043	4.291	***	0.190	Established
H5: WTP→TBP	0.406	0.057	7.094	***	0.387	Established

*and *** denote significance at the 10% and 1% levels, respectively.

Based on the standardized results, ecoliteracy has a positive effect on increasing farmers’ WTP in PPW governance. First, ecological cognition and ecological knowledge and skills have a similar degree of effect on farmers’ WTP in governance, and farmers’ behaviors are affected by both in a balanced way, but the facilitating effect of ecological knowledge and skills is stronger than ecological cognition. Farmers make rational judgments about the work and effects of PPW governance. When farmers believe that participation in PPW governance is beneficial to the rural environment, and when individuals believe that time and labor costs are conducive to the promotion of pesticide packaging recycling and disposal, they will be able to make a positive judgment and be more willing to participate in PPW management. Meanwhile, an improvement in the knowledge and skills of farmers can directly and effectively enhance their WTP in PPW management. Ecological knowledge refers to people’s basic understanding of the environment and environment-related issues. When farmers have PPW recycling and disposal processes, PPW policy and other knowledge reserves, and a sense of PPW governance identity, it can promote farmers’ WTP in PPW governance. Second, ecological emotions are slightly higher than ecological values regarding farmers’ WTP in PPW governance, but both are lower than ecological cognition and ecological knowledge and skills. Owing to the quasi-public good attributes of the rural environment, farmers generally believe that the responsibility for PPW governance should be borne by the government and pesticide companies, and there is a wide range of “free-riding” behavior. With the improvement of farmers’ cognition and knowledge and skills, they gradually form ecological emotions and values; this subjective perception helps induce farmers to participate in PPW management.

There are two effective paths for ecoliteracy’s effect on farmers’ participation in PPW governance: ecological cognition and ecological knowledge and skills. Ecological emotions and ecological values, however, do not have a significant effect on farmers’ participation in PPW management. A possible reason is that the observed variables in the latent variable “ecological emotion” include the sense of governance responsibility, pleasure, and guilt, and the observed variables in “ecological values” include rural environmental interests, ecological protection values, and farmers’ behavioral values, which are mainly subjective volitional conditions and are more psychological. These variables are not sufficient to contribute to positive or negative PPW participation behavior. At the same time, the establishment and cultivation of farmers’ ecological emotions and values are long-term, continuous processes, in which farmers first gradually form the antecedent of participation in PPW governance—that is, WTP in PPW governance. Meanwhile, the ecological knowledge and skills of farmers can be rapidly improved in a short period through learning and training, thus directly regulating their participation in PPW governance behavior. Farmers with rich knowledge and skills have a higher degree of participation in PPW management, thus reducing “throw-away” behavior and implementing PPW recycling and disposal. If they lack ecological knowledge and skills, they will not be able to implement pollution reduction and environmental improvement behaviors. “How-to” knowledge drives farmers to improve their “can-do” practices to achieve the desired goals. Farmers consider not only economic benefits when making production decisions but also the ecological effects of specific actions. This will lead to the better implementation of PPW governance behaviors and make it possible for them to cooperate with the government and other actors, thus reducing barriers to the implementation of behavior.

Mediation effect test

Based on the results of the hypothesis testing, there are four mediating paths between farmers’ WTP in governance and “ecoliteracy–participation in PPW governance behavior”: (1) ecological cognition → WTP in governance → participation in PPW governance behavior; (2) ecological emotion → WTP in governance → participation in PPW governance behavior; (3) ecological values → WTP in governance → participation in PPW governance behavior; and (4) ecological knowledge and skills → WTP in governance → participation in governance PPW behavior. Among them, the direct effects of ecological emotions → participation in PPW governance behavior, and ecological values → participation in PPW governance behavior are not significant, and therefore, the mediating effect is a complete mediating effect. Combined with the hypothesis testing results to further analyze the mediating effect, bootstrapping is used to test the model, and the mediating effect of farmers’ WTP in PPW governance on ecoliteracy’s effect on participation in PPW governance behaviors is examined under conditions of 1000 draws of the sample and the set conditions of the 95% confidence interval. Referring to Mackinnon⁸⁰, we experimentally derive the optimal method of bias correction in the non-parametric bootstrap method to correct the bias in the confidence intervals. Table 5 shows that the confidence intervals of the bootstrap test for the four mediating paths do not contain 0. This indicates that farmers’ WTP in PPW governance plays a mediating role between ecoliteracy and participation in PPW governance behaviors. However, there are differences in the magnitude of the direct and indirect effects of the dimensions on the participation of farmers in the governance of PPW behavior. Nevertheless, they all confirm that farmers’ WTP in PPW governance plays a mediating role in the influence of ecoliteracy on participation in PPW management.

Table 5.

Path parameter estimation results of the mediating effect test.

Research hypothesis: Mediation path	Indirect effect	Direct effect	Total effect	LLCL	ULCL	Mediating effect
H5a: EC→WTP→TBP	0.022	0.052	0.240	0.078	0.166	Established
H5b: EE→WTP→TBP	0.018	−0.008	0.199	0.055	0.126	Established
H5c: EV→→WTP→TBP	0.015	−0.074	0.035	0.023	0.084	Established
H5d: ES→→WTP→TBP	0.024	0.084	0.275	0.090	0.184	Established

LLCL is the lower limit of the 95% confidence interval; ULCI is the upper limit of the 95% confidence interval.

Heterogeneity test

Above, the effects of ecological on farmers’ PPW governance behavior, i.e., the homogeneous treatment effect, have been tested at the full sample level, and the heterogeneity of their effects has not yet been considered. There are certain differences in the endowment of farmers and their geographical location, which, if not differentiated, are likely to mask the true effect of ecological on farmers’ participation in pesticide packaging waste management behavior. Considering the differences in group characteristics, this paper explores the heterogeneity of the effects of ecological on farmers’ PPW governance behavior across different planting scales, planting categories, and areas supported by government programmers. It has been well-documented that farm household heterogeneity is the micro cause of differences in ecological conservation behavior, as shown in Table 6, and scholars have carried out extensive micro analyses based on the heterogeneous farm household model.

Table 6.

References mentioned in the response.

Chapters	Content modification	Additional references
Grouping of planting categories	Established research have confirmed that differences in planting categories also lead to different decision-making by farmers, and that different planting categories and ecoliteracy influence farmers’ participation in PPW governance behavior.	Freeman et al(2003). Fertilizer use in semi-arid areas of Kenya: analysis of smallholder farmers” adoption behavior under liberalized markets.
Planting scale grouping	Research findings have confirmed that planting size is an important factor influencing farmers’ implementation of environmentally friendly production behaviors, that size heterogeneity influences farmers to produce different behavioral decisions, and that large-scale households strengthen the impact of ecoliteracy on their implementation of environmental behaviors compared to small-scale farmers.	Ren et al(2022). Uncovering the Deviation of Farmers’ Green Manure Planting Willingness and Behavior. Kong et al(2019). Differential Analysis of Environmentally Friendly Production Behavior of Farming Households of Different Scales - Based on Research Data of 1059 Farming Households in 7 Provinces of China.
Access to government project support subgroup	It has been confirmed that whether or not the farmers are in the government project support area shows obvious heterogeneous characteristics, the farmers in the government project support area have higher ecoliteracy, and the influence effect of the farmers’ behavioral change is gradually highlighted.	Li et al(2022). Impacts of Risk Perception and Environmental Regulation on Farmers’ Sustainable Behaviors of Agricultural Green Production in China.

Grouping of planting categories

Established research have confirmed that differences in planting categories also lead to different decision-making by farmers⁸¹, and that different planting categories and ecoliteracy influence farmers’ participation in PPW governance behavior. Based on Eqs. (1) and (2), we use multiclause validated factor analysis to test the measurement equivalence of ecoliteracy among different categories of growers. Regarding cropping structure, it can be divided into three groups according to crop variety: grain, economic, and mixed grain and economic growers⁸². Following the procedure for testing the measurement constancy of a model in the studies of Wang⁸³ the measurement equivalence of ecoliteracy’s effect on the participation of farmers in PPW governance behaviors consists of the following five steps. (1) Establish a single-group baseline model, and perform a validation factor analysis of the factor structure of each factor structure in the SEM separately to establish an acceptable fitted single-group baseline model. (2) Test the morphological equivalence model (M1). This tests whether factor structures such as the number of factors and the factor path model settings are equivalent between different groups and also tests the cross-group equivalence of different groups. (3) Test the measurement of weak equivalence (M2). This adds the factor loading equivalence constraints on the basis of the M1 morphologically equivalent model and tests whether the factor loadings are the same between different groups. (4) Test for strong equivalence (M3). Add factor variance–covariance equivalence restrictions based on M2 weak equivalence to test whether the intercepts of each observed variable are equal between different clusters. (5) Test the residual equivalence model (M4). Add observed variable equivalence restrictions based on M3 strong equivalence to test whether the residuals of each observed variable are equal between different clusters. Tables S5 show the model fit indices. The results of the single-group validated factor analyses show that the data from the baseline model fit well, with p-values greater than 0.05 for both M2 and M4. Although the p-value for M3 is less than 0.05, it complies with measurement equivalence since ΔTLI < 0.05 and ΔCFI ≤ 0.01⁸⁴, which are consistent with the measurement equivalence test. This indicates that ecoliteracy has strict equivalence across the planting categories for farmers’ participation in PPW governance behavior—that is, there is consistency in the factor structure, factor loadings, intercepts, and residuals across items.

Table S5 depicts the overall constancy of the model by constructing the nested model step by step from the baseline model with different sample SEMs. The indices of model fitness for each cluster are at the desired level; thus, cluster analysis can be performed on the basis of the original model. To verify whether the research hypothesis holds for different agricultural growers, and to refine the intrinsic relationship between the subject’s behavior and other latent variables, Table 7 show the results for the model path parameters based on cultivation category. In the cluster analysis of farmers’ participation in PPW governance behavior, there are significant differences in ecoliteracy’s effect on participation in PPW governance behavior among grain growers, economic growers, and mixed grain and economic growers.

Table 7.

Path parameter estimation results of the model heterogeneity test.

Research hypothesis: influence path	Food category	Economic category	Mixed grain and economy
H1a: EC→WTP	0.266***	0.320***	0.307***
H1b: EC→TBP	0.221**	0.038	0.136*
H2a: EE→WTP	0.168*	0.238***	0.218***
H2b: EE→TBP	−0.022	0.139*	0.091
H3a: EV→WTP	0.129*	0.219***	0.009
H3b: EV→TBP	0.050	−0.007	−0.017
H4a: ES→WTP	0.389***	0.238***	0.442***
H4b: ES→TBP	0.191*	0.161*	0.181*
H5: WTP→TBP	0.387***	0.269**	0.457***

*, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.

The effect of ecoliteracy on farmers’ participation in PPW governance behavior varies by cultivation category. First, from the test results, the seven hypotheses in the coefficient test of the structural equation simulation path for grain farmers are all valid and consistent with the parameter estimation results of the overall SEM of the effect paths, which further validates the findings. Second, the path coefficients of “ecological cognition → participation in PPW governance behaviors” are not significant, but the path coefficients of “ecological emotion → participation in PPW governance behaviors” show a significant effect. This could be because economic crops such as vegetables, traditional Chinese medicinal herbs, and melons place higher demands on the land environment and soil quality, farmers have stronger functional and emotional attachments to land, and a strong sense of responsibility has been developed through long-term agricultural production. Thus, economic crop growers have a stronger sense of both pleasure and guilt resulting from their participation in PPW governance, and ecological emotions have a positive effect on promoting governance behaviors. Third, the effects of ecological cognition, ecological emotion, and ecological knowledge and skills on WTP in governance are confirmed by the mixed grain and economic farmers. This also reflects the significant effect of ecological cognition, ecological knowledge and skills, and WTP in governance on participation in governance behavior. Notably, the coefficient of the path of “ecological values → participation in PPW governance behavior” is not significant, rejecting H3b. This could be because compared with economic and grain growers, mixed crop farmers have more time investment and labor intensity, less remaining leisure time, and less attention to the village environment. Thus, ecological values do not affect participation in governance behaviors.

In terms of the significance of the effect, the significance of the path relationship of mixed grain and economic crop growers is overall higher than that of economic crop growers and grain crop growers. Among the effects of the endogenous latent variables, farmers’ WTP in governance is more significant in the mixed grain and economic groups than in the grain and economic groups, and the coefficients of the path relationships are higher in all three groups than in the overall model. In terms of the influence of exogenous latent variables on endogenous latent variables, farmers in the mixed grain and economic group pay attention to the influence of knowledge skills and ecological cognition on PPW management WTP and behavior. Compared with farmers in the mixed grain and economic groups, the food category group and economic category group emphasize the role of ecological cognition and ecological knowledge and skills; however, farmers in the economic group are more inclined to adopt PPW governance behaviors based on the cultivation of ecological emotions.

Planting scale grouping

Research findings have confirmed that planting size is an important factor influencing farmers’ implementation of environmentally friendly production behaviors⁸⁵, that size heterogeneity influences farmers to produce different behavioral decisions⁸⁶, and that large-scale households strengthen the impact of ecoliteracy on their implementation of environmental behaviors compared to small-scale farmers. Referring to the division of cultivation scale in the studies of Qin and Lu⁸⁷, we divide the research area into three groups: small scale < 3 mu, 3 mu < medium scale < 10 mu, and large scale > 10 mu. We test the morphological equivalence model (M1), weak equivalence model (M2), strong equivalence model (M3), and residual equivalence model (M4). Table S6 shows the estimation results. Although the p-value of M1–M4 is less than 0.05, the two indicators of ΔTLI and ΔCFI are less than 0.05 and less than or equal to 0.01, respectively, which is in line with the threshold value of the measurement equivalence test. This indicates that the multicluster analysis model fits well with the sample data, and ecoliteracy has strict equivalence at different planting scales in its effect on farmers’ participation in PPW governance behavior.

In the multicluster analysis of farmers at different planting scales, small-, medium-, and large-scale farmers show obvious differences in ecoliteracy’s effect on their participation in PPW treatment behavior. Regarding the influence path of ecoliteracy’s effect on governance WTP and behavior, Table 8 shows that both medium- and large-scale farmers pass the significance test, thus verifying hypotheses H1a, H1b, H2a, H3a, H4a, H4b, and H5. The influence of small-scale farmers’ ecoliteracy on their governance WTP and behavior does not fully verify the hypotheses. Ecological cognition → participation in governance behavior” and “ecological values → participation in governance willingness” do not pass the test; thus, H1b and H3a are not confirmed.

Table 8.

Path parameter estimation results of the model heterogeneity test.

Research hypothesis: influence path	Small scale	Medium scale	Large scale
H1a: EC→WTP	0.413***	0.218***	0.245**
H1b: EC→TBP	0.091	0.131*	0.175*
H2a: EE→WTP	0.203***	0.264***	0.179*
H2b: EE→TBP	0.072	0.096	0.04
H3a: EV→WTP	0.080	0.128*	0.163*
H3b: EV→TBP	−0.036	0.089	−0.057
H4a: ES→WTP	0.294***	0.408***	0.377***
H4b: ES→TBP	0.166*	0.183*	0.194*
H5: WTP→TBP	0.430***	0.307***	0.429***

*, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.

There are obvious differences in the intensity and significance levels of the variables affecting behavior among farmers at different scales. First, for small-scale farmers, the order of influence on farmers’ WTP in governance from strong to weak is ecological cognition, ecological knowledge and skills, and ecological emotion. The order of influence on farmers’ participation in PPW governance behavior is farmers’ WTP in PPW governance and ecological knowledge and skills. Different from the ungrouped model, “ecological cognition → participation in PPW governance behavior” and “ecological values → WTP in PPW governance” are not verified. The main reason is that small-scale farmers have a high degree of land fragmentation and a small average application of pesticides per mu. Although they believe the recycling and treatment of PPW significantly affect the rural environment, it is difficult for ecological cognition to rise to ecological values, and their wishes and behaviors are contrary. We also find that the plots of small-scale farmers are scattered. For example, the plots in Nanzhuang Village, Jize County, are trivial, with per capita land plots of 5–7. In particular, grain growers are scattered because of the small scale of the plots; thus, they often choose to package and landfill waste pesticides and discard them at will. Second, regarding medium-scale farmers, the influence order of farmers’ WTP in governance is ecological knowledge and skills, ecological emotion, ecological cognition, and ecological values. The order affecting farmers’ participation in governance behavior is WTP in governance, ecological cognition, and ecological knowledge and skills. The influence coefficient of “WTP in PPW governance → participation in PPW governance behavior” is 0.307. Ecoliteracy promotes the generation of farmers’ WTP in PPW governance, and finally, it is most effectively transformed into governance behavior. Third, among large-scale farmers, ecological knowledge and skills, ecological cognition, ecological emotion, and ecological values promote their WTP in governance. WTP in governance is further transformed into governance behavior, and the parameter estimation result for this path is as high as 0.429. In addition, the effect coefficient of “ecological knowledge and skills → participation in PPW governance behavior” is 0.194, which exceeds 0.183 for medium-sized farmers and 0.166 for small-scale farmers. Large-scale farmers pay attention to policies related to benefiting farmers and environmental protection. Comparative incomes from agriculture are high, the demand for farmland environmental governance is high, and farmers’ “self-initiative,” or willingness to take action, is gradually enhanced.

Access to government project support subgroup

It has been confirmed that whether or not the farmers are in the government project support area shows obvious heterogeneous characteristics⁸⁸, the farmers in the government project support area have higher eco-literacy, and the influence effect of the farmers’ behavioral change is gradually highlighted. Based on the field investigation, all samples are divided into five groups: “provincial pilot project of PPW disposal,” “municipal pilot project of PPW disposal,” “beautiful countryside construction project,” “beautiful countryside construction project and industrial poverty alleviation project,” and “no project support.” Table S7 shows the test results for the measurement equivalence of different government projects. In the morphological equivalence test, CFI = 0.92, TLI ≥ 0.91, RSMEA ≤ 0.07, and SRMR ≤ 0.05. The model fit index meets the requirements of surveying, indicating that morphological equivalence is established and that the next equivalence analysis can be performed. Then, in the weak equivalence M2 test, according to the results of the model fitting index, the CFI and TLI values are less than 0.010, showing that the index factor load is equivalent across groups; that is, weak equivalence holds. Then, in the strong equivalence test, the CFI and TLI values are − 0.026 and − 0.024, respectively, which are less than 0.010, indicating that the intercept equivalence of different groups is valid—that is, strong equivalence. In the final strict equivalence test, the CFI and TLI values are both less than 0.010, and the strict equivalence test is also valid.

As shown in Table 9, in the path parameter results of the heterogeneity test of different government project support models, there are great differences in the degree of ecoliteracy’s effect on farmers between those with government support and those without government project support. In the influence path of “ecoliteracy → WTP in governance,” the provincial pilot project of PPW recycling and disposal and not supported by government projects verifies hypotheses H1a, H2a, H3a, and H4a. The municipal pilot project team of PPW recycling and treatment verifies H1a and H2a. The beautiful countryside construction project team verifies H1a and H4a. The industrial poverty alleviation and beautiful countryside construction project groups verify H2a and H3a. In the influence path of “ecoliteracy → participation in PPW governance behavior,”—a non-government support project team—only H4b is supported. In the influence path of “WTP in governance → participation in PPW governance behavior,” the provincial pilot project of PPW recycling, beautiful countryside construction project, and farmers’ WTP in governance without government project support group have a significant positive effect on behavior.

Table 9.

Path parameter estimation results of the model heterogeneity test.

Research hypothesis: influence path	Provincial pilot project of PPW recycling and disposal	Municipal-level pilot project of PPW recycling and disposal	Beautiful countryside construction project	Industrial poverty alleviation and beautiful countryside construction project	Not supported by government projects
H1a: EC→WTP	0.212***	0.339*	0.424***	0.167	0.322***
H1b: EC→TBP	0.077	0.001	0.183	0.042	−0.039
H2a: EE→WTP	0.212***	0.419*	0.035	0.291*	0.184**
H2b: EE→TBP	0.090	0.211	0.088	0.118	0.09
H3a: EV→WTP	0.120*	0.256	−0.067	0.385**	0.145*
H3b: EV→TBP	0.022	−0.068	0.107	0.100	0.112
H4a: ES→WTP	0.335***	0.033	0.600***	0.179	0.349***
H4b: ES→TBP	0.020	0.405	−0.006	0.290	0.232*
H5: WTP→TBP	0.370***	0.326*	0.575*	0.017	0.542***

*, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.

The degree of ecoliteracy’s effect on farmers’ participation in PPW governance behavior is quite different in the heterogeneity test involving different government project support. Regarding the influence path of “ecoliteracy → WTP in governance,” farmers’ ecological cognition, ecological emotions, ecological values, and ecological knowledge and skills have a significant positive effect on their WTP in governance, and farmers can effectively transform their WTP in governance into behavior. In the provincial pilot project of PPW recycling, farmers’ WTP in governance is influenced by ecological cognition and ecological emotion. The parameter of ecological cognition to WTP in governance in beautiful countryside construction projects is as high as 0.424. For industrial poverty alleviation and beautiful countryside construction project teams, ecological values and ecological emotions play a significant role in affecting farmers’ WTP in governance. Without the government project support group, WTP in PPW governance is affected by ecological cognition, ecological emotions, and ecological values. Regarding the influence path of “WTP in governance → participation in PPW governance behavior,” the estimation results for the path parameters of provincial pilot projects, beautiful countryside construction projects, and projects without government support are significant, verifying hypothesis H5. For the influence path of “ecoliteracy → participation in governance behavior,” only ecological knowledge and skills in non-government-supported project groups significantly affect governance behavior, while the influence path of ecoliteracy on participation behavior in other groups does not pass the test. A possible reason why most influencing factors are not significant is that although ecoliteracy forms the internal influencing mechanism of farmers’ WTP in governance, environmental regulation, as an external factor, can effectively reduce the possibility of a paradox between behavior and willingness.

Discussion

Further interpretations

Our results reveal the influence of ecoliteracy on farmers’ participation in PPW governance behavior. We extend the study in terms of farmers’ heterogeneity, analyzing the mechanism of farmers’ participation in PPW governance behavior under different planting types, at different planting scales, and for different areas supported by different governmental projects. As direct participants in pesticide use and PPW recycling, the behaviors of farmers reflect the economic rationality of profit maximization, as well as the ecological rationality of “altruism” and “collective action logic⁸⁹”. The path of ecoliteracy is “willingness–behavior,” and the hierarchy of ecoliteracy is as follows: ecological cognition is the foundation and premise for forming ecoliteracy, ecological emotion is the expression of ecoliteracy and the tendency toward responsibility, ecological value is the sublimation of the perception of the environment and the sense of responsibility, ecological knowledge and skills are the vision of ecoliteracy, and ecological knowledge and skills are the vision of ecoliteracy. Ecological cognition, ecological emotion, and ecological values subjectively contribute to the internal orientation of farmers; ecological knowledge and skills objectively contribute to farmers’ communication and learning mechanisms. Ecoliteracy can drive small farmers’ enthusiasm to participate in PPW management from the internal psychological level to shift farmers from WTP to actual behavior.

Ecoliteracy has a significant positive effect on farmers’ participation in PPW governance behavior, consistent with Sun⁹⁰, Wang⁹¹, and Lu⁹², the logic of farmers’ participation in PPW governance behavior is consistent with the theory of farmers’ behaviors. There are three logical lines: “ecoliteracy → farmers’ WTP in governance,” “ecoliteracy → farmers’ participation in PPW governance behavior,” and “farmers’ WTP in governance → participation in PPW governance behavior.” According to Table 2, ecological cognition, ecological emotion, ecological values, and ecological knowledge and skills have a significant positive effect on farmers’ WTP in governance. Ecological cognition, ecological knowledge and skills, and WTP in governance have a positive effect on the response to participation in PPW governance behavior. WTP in governance significantly affects governance participation behavior at the 1% statistical level, and the standardized coefficient of this path is as high as 0.387. This further indicates that WTP in PPW governance significantly affects farmers’ participation in PPW governance. This further indicates that farmers’ WTP in PPW governance has a “spontaneous” effect that dominates the behavioral response; thus, H1, H2a, H3a, H4, and H5 can be verified. In addition, the “knowledge spillover” effects of ecological awareness and ecological knowledge and skills play an important role in driving farmers’ participation in governance behaviors. Therefore, farmers’ participation in PPW governance is an organic combination of “spontaneity” generated by the subjective logic of ecological cognition, ecological emotion, ecological values, and “knowledge spillover” generated by the objective logic of ecological knowledge and skills.

Another important research conclusion is that farmers’ WTP in PPW governance has a mediating effect between ecoliteracy and participation in management behavior. Further confirming the conclusions of Lu⁹³, Jiang⁹⁴, and Zhao⁹⁵, according to the total effect value estimates in descending order, the influence paths are “ecological knowledge and skills → WTP in governance → participation in PPW governance behavior,” “ecological cognition → WTP in governance → participation in PPW governance behavior,” “ecological emotion → WTP in governance → participation in PPW governance behavior,” “ecological values → WTP in governance → participation in PPW governance behavior.” The findings shown in Table 3 indicate that improving farmers’ behavioral implementation can start with ecological knowledge and skills and ecological cognition.

For different types of farmers, there are differences in ecoliteracy’s effect on their behavioral responses to participation in PPW governance. The validated factor analyses in Tables 4, 5 and 6 show that the hypothesis of measurement equivalence among farmers in different planting categories, different planting sizes, and different government-supported project subscales is valid. That is, the subscales have equal significance in different groups, and the heterogeneity of farmers has different paths in the responses to ecoliteracy and participation in PPW governance behavior. First, the degree of ecoliteracy’s effect on participation in PPW governance varies across farming categories. Regarding the degree of influence of each dimension of ecoliteracy on farmers’ participation in governance behavior, the mixed grain and economic group has more influence than the economic group and the grain group. Regarding the significance of the path parameter estimates and the size of the standardized coefficients, ecological knowledge and skills among farmers in the grain category play a dominant role in WTP in governance. The ecological knowledge of farmers in the economic category has the most influence on WTP in governance, and the ecological cognition of the mixed grain and economic group has the highest degree of influence on WTP in governance. The path of the effect of WTP in governance participation behavior occupies a central position. The results of the path parameter of the model heterogeneity test of the mixed grain and economic crop farmers are higher than those of the grain and economic crop farmers as a whole. The more that farmers can conclude it is relatively easy to participate in PPW governance based on their own ecoliteracy, the more they tend to have a positive attitude toward PPW governance, and the greater the tendency to respond positively. Second, the ecoliteracy of farmers at different planting scales has different degrees of influence on participation in PPW governance. The participation behaviors of farmers in the large-scale group are stronger than those of farmers in the medium- and small-scale groups. Farmers in the large-scale group have higher ecoliteracy, and WTP in governance strengthens the positive effect of ecoliteracy. The ecological knowledge and skills of farmers in the medium-sized group play a more significant role; The ecological knowledge of small-scale farmers dominates their WTP in PPW management, which is then converted into behavior. Third, the ecoliteracy of farmers in different government project support areas has different degrees of influence on their participation in PPW governance. Farmers in areas that receive government project support have more willingness and prefer to implement governance behaviors. In the PPW return treatment provincial pilot project and beautiful countryside construction project groups, ecological knowledge and skills, ecological cognition, and WTP in governance jointly drive farmers to participate in governance behaviors. In the PPW return treatment municipal pilot project group, farmers are affected by ecological cognition and ecological emotional strength. In the industrial poverty alleviation and beautiful countryside construction project groups, ecological values dominate. For areas that do not receive government project support, WTP in governance is the main influencing factor. However, in the project subgroups that do not receive government support, most of the influence paths are unclear, providing a possible direction for future research to further explore the mechanism of environmental regulation’s influence on farmers’ participation in PPW governance behavior.

Policy recommendations

Farmers’ ecological is positively related to PPW governance effectiveness. In addition to environmental regulation, ecoliteracy can also influence farmers’ behaviors through the internal drive path, forming an endogenous, long-term mechanism for farmers to participate in PPW governance. China’s promotion of PPW governance should therefore focus on enhancing the value orientation of ecoliteracy and stimulate the endogenous motivation of farmers to participate in governance. (1) Enhance farmers’ ecological awareness. Ecological awareness is the basis for ecoliteracy formation and should be carried out in two aspects. First, create a collaborative governance environment, issue policy brochures, organize on-site exchanges, improve publicity on safe PPW disposal and governance policy, improve farmers’ public literacy, enhance the sense of honor of PPW governance, and promote the generation of collective action. Second, improve the ecological awareness of small-scale farmers. Small-scale farmers have the highest level of ecological cognitive influence, and a village environmental education system should be built. Use the village party member and cadre training base, legal propaganda, and education training base to popularize the agricultural origin of environmental protection, green agricultural production, and agricultural waste management to enhance ecological cognition and cultivate the idea of synergy between farmers’ agricultural production and operation and rural environmental protection. (2) Offer in-depth ecological knowledge training and technical guidance for farmers. The accumulation of ecological knowledge and skills and technology diffusion can promote farmers’ participation in PPW governance. Broaden the channels for farmers to obtain green production information. Use digital technology to promote the government’s ecological governance work, extend government services to the grassroots, and use social media and other platforms to promote safe pesticide use, the scientific application of production materials, and waste management. In addition, make use of “government + agricultural research institutes,” technical cooperation, and field technical guidance or demonstration. In addition, use technical cooperation between the government and agricultural research institutes to organize regular technical guidance in the field or demonstration plots to enhance farmers’ confidence in engaging in eco-agriculture. Support agricultural technology training. Improve the technical service system for farmland ecological protection; strengthen technologies such as those for soil improvement and restoration, PPW governance, and resource utilization; and improve farmers’ skills in dealing with farmland ecological problems. (3) Cultivate farmers’ emotions and values. Values are an important parameter in the collaborative management of PPW and play a leading role in farmers’ participation in PPW management; thus, it is particularly important to reshape consensus on values. Pilot counties, as the basis for the collaborative construction of pesticide waste management public service models, prompt farmers to take the initiative to integrate PPW governance and enhance brand protection awareness from “embedded” to “symbiosis.”

Limitations and future research directions

This study has some limitations. For example, the field data are cross-sectional, and given the difficulty of obtaining panel data, the trend study of changes in the PPW governance behaviors of farmers is also somewhat lacking. Further, the analysis of the heterogeneity of farmers still needs to be refined. However, owing to the limitation of sample size, the heterogeneity of farm households could be further subdivided. In the future, the research area can be further expanded to fit the characteristics of agricultural production and operation in each region, and the driving mechanism of farmers’ participation in PPW governance behavior can then be further explored.

Conclusions

This study uses SEM to analyze ecoliteracy’s effect on farmers’ participation in PPW governance behavior using data from 1118 farmers in Cangzhou, Baoding, Tangshan, Zhangjiakou, and Handan in Hebei Province. We use a multicluster structural model to test the heterogeneity according to planting category, planting scale, and government project support for farmers. The conclusions are as follows: Ecoliteracy has a significant positive effect on farmers’ participation in PPW governance behavior, and farmers’ WTP in governance has a mediating effect between ecoliteracy and governance participation behavior. (1) Ecological knowledge and skills and ecological cognition are important factors that enhance farmers’ WTP in governance. The richer the farmers’ knowledge reserve, the more ecological cognition they have, the more positive their attitude toward environmental protection, and the more they can mobilize their subjective initiative. Ecological emotion plays a stronger role than ecological values in farmers’ WTP in governance, but the degree of influence is lower than that of ecological knowledge and skills and ecological cognition. In addition, there is a progressive relationship among ecological cognition, ecological emotion, and ecological values, and ecological cognition plays a fundamental role in farmers’ behaviors and eventually generates ecological values. (2) Ecological cognition and ecological knowledge and skills have a positive effect on farmers’ participation in governance behavior. In the short term, farmers can significantly improve their ecological knowledge and skills through learning and training, which directly affect their behavior. (3) The effect of WTP in governance is different in each mediating pathway. The total effect estimates are ranked as ecological knowledge and skills, ecological cognition, ecological emotion, and ecological values. This suggests that if we want to improve the degree of farmers’ participation in governance behavior, we can start with ecological knowledge and skills and ecological cognition. (4) Heterogeneity exists in the PPW governance behaviors of farmers at different planting scales, in different planting categories, and in different government project support areas. Mixed grain and economic crop growers pay attention to the acquisition of knowledge and skills and ecological cognition to improve WTP in PPW management and the effect of behavior. Economic crop growers are inclined to implement PPW management based on the cultivation of ecological emotions. Large-scale farmers and those in areas that receive support from the government are likely to implement the management of behavior. Environmental regulation effectively reduces the possibility of behavior and the willingness to contradict each other.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Yang Song: Conceptualization, investigation, methodology, writing—original draft preparation. Hai-Xia Cui: Methodology, Writing—review& editing. Yi-Xiang Zong: Conceptualization, methodology, visualization, formal analysis, writing—review& editing. Shi Yin: Conceptualization, Writing—review & editing, Supervision, Validation. All authors approved the manuscript and give their consent for submission and publication.

Funding

The funding was supported by Funding for Key Research Bases of Humanities and Social Sciences in Higher Educational Institutions in Hebei Province, Funded by Science Research Project of Hebei Education Department “Collaboration Mode and Benefit Distribution of Vegetable Supply Chain in Hebei Province to Beijing (JCZX2024008)”, National Social Science Foundation of China(23BGL266), Hebei Agriculture Research System (HBCT2023100301), Research Centre for Rural Revitalization Strategy of Hebei Province(HB21ZK11).

Data availability

Due to data protection and participant confidentiality concerns, datasets generated during and/or analyzed during the current study are available from the corresponding author upon request.

Declarations

Competing interests

The authors declare no competing interests.

Statement

All methods were carried out in accordance with relevant guidelines and regulations. All experimental protocols were approved by College of Economics and Management, Hebei Agricultural University. Institutional informed consent was obtained from all subjects (farmers).