A study of intention to recycle pesticide packaging among rural farmers in China based on the logistic model

A consequence of the improvement in China’s agricultural production capacity has been increased production of pesticide packaging waste, which is a major obstacle to the sustainable development of rural areas. Although the Chinese government has made great efforts to control pollution by waste pesticide packaging, a unified recycling system has not yet been established. Therefore, determining whether farmers are willing to participate in recycling of pesticide packaging is of great relevance. This study conducted a random questionnaire survey in which 459 responses were obtained in rural areas in Liaoning, Henan, Guizhou, Anhui, Gansu, and Hunan provinces. The disorderly classification logistic model was used to analyze the individual characteristics of farmers, pesticide container sizes, perception of risk, conditions of local facilities, and awareness. The result showed that the individual characteristics of farmers and awareness had no influence of their willingness to participate in recycling, whereas container size, perception of risk, and condition of local facilities did have an influence. Container specifications have the strongest correlation with farmers’ willingness to recycle, and attempts can be made to promote pesticide packaging of about 500 ml; raising farmers’ awareness of environmental safety and possible health hazards, all of which can increase farmers’ participation in the work. Using these results, this study proposed suggestions for pesticide container waste disposal. Attempts can be made to establish a suitable recycling model and management mechanism through the establishment of additional recycling points. Refine pesticide packaging waste recycling, collection and storage, disposal and other measures to purify the environment of origin, to achieve green agricultural development, and promote the construction of ecological civilization.


Introduction
Large-scale agricultural production in China has increased significantly. Pesticides are widely used in agricultural production in China to control pests and improve crop yield (Sharma et al 2020). However, a consequence of pesticide use has been the production of pesticide packaging waste. This waste places a large burden on the environment and hinders sustainable development of rural areas. Pesticides have been ranked with heavy metals, radioactive nuclear elements, and persistent organic pollutants as the most damaging pollutants to human health (Picuno et al 2020). The pesticide packaging is typically made of plastic, glass, metal, chemical fibers, and some other materials (Li and Huang 2018). Items made of this material do not degrade easily and can seriously affect the health quality of the soil (Jin et al 2018). Some pesticide residues in packaging can also be washed off by rainfall into surface and groundwater resources. Long-term exposure to these pollutants will cause them to accumulate in the human body, resulting in harmful effects on growth, development and metabolism, which can lead to serious cardiovascular and cerebrovascular diseases, etc (Sharma et al 2020). Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
There is increasing recognition for the need for improvements to food safety and conservation of the ecological environment. Consequently, many countries have been focusing increasing attention to the recycling and disposal of pesticide packaging. Germany issued the Packaging Waste Treatment Regulations of 1991 (He and Jin 2013), and also established the PAMIRA Institution in 1996 for the recycling of pesticide packaging waste (Li et al 2015). Statistical records indicate that 78% of all pesticide packaging was recovered in 2017 (Liu et al 2021). The Brazilian National Pesticide Industry Association began to advocate the management of pesticide packaging waste in 1993. The National Institute of Empty Container Treatment (inpEV) was established in 2001, which obtained agreement among 99% of pesticide manufacturers, dealers, and farmers for the disposal of pesticide packaging waste (Zhang 2014). The rate of recovery of waste pesticide packaging in Brazil in 2005 reached 87% (Xu and Zhang 2021). The Canadian Pesticide Industry Association has proposed the Stewardship First project for the recycling of pesticide packaging waste, and in 2011 the recycling rate of pesticide packaging in Canada has reached 73% (Liu et al 2023).
Although research on pesticide packaging waste in China started relatively late, some provinces in China have developed mature recycling protocols. The characteristics of agricultural production in different areas in Beijing can be divided into four modes of recycling: (1) barter; (2) cash payback; (3) recycling contracts, and; (4) entrusting a third party (Wang et al 2018). Local governments at the street or town level in Xiqing District in Tianjin are responsible for establishing points for the paid recovery of pesticide packaging (Feng et al 2020). Guangdong Province (Xu et al 2023) through a combination of incentive model and agreement recycling model, Heilongjiang Province (Li et al 2022a(Li et al , 2022b through digital management, innovative recycling model are to achieve the resourceful reuse of pesticide packaging waste. The effective recovery of pesticide packaging can promote ecological and human health. However, the survey found that nearly 62% of Chinese farmers would discard pesticide packaging directly into the water or surrounding environment (Wei and Jin 2014). The degree of recycling of pesticide packaging is closely related to the current individual characteristics and family status of farmers. The majority of the elderly in rural areas, serious aging, relatively low literacy, and the lack of labor in each household are several internal factors that can affect the recycling of pesticide packaging (Li et al 2020). Economic costs are the main external factor preventing farmers from transforming environmental awareness into environmentally friendly behavior (Wei and Du 2018). In response to this phenomenon, some regions have taken relevant measures. For example, Zhejiang Province provides financial subsidies to low-income, large-scale farmers to incentivize pesticide packaging recycling behavior (Li et al 2022a(Li et al , 2022b. The probability of using the housing loan rebate policy to incentivize farmers to recycle in Jiangsu Province increased by 44.8% (Hu et al 2022). Guangdong Province has constructed three recycling models applicable to the region, namely agricultural stores, cleaning teams or rural enterprises unified recycling (Cai 2019). However, the above survey results and related policies are only applicable to the local situation in a certain province or a certain region, and there is no overall survey for the whole China, so as to come up with a recovery system and policy suitable for the whole country.
Therefore, based on the results of previous studies, this study aimed to identify five aspects of farmers' willingness to participate in pesticide packaging waste recycling: (1) farmers' personal characteristics; (2) pesticide container specifications; (3) risk perception; (4) local facilities, and; (5) awareness, based on a disordered multicategorical logistic model. The willingness of farmers to participate in recycling efforts and related influencing factors are explored. The problems in the process of pesticide packaging waste disposal were analyzed, and feasible countermeasures were proposed to further strengthen the recycling of pesticide packaging waste in the future. Feasible countermeasure suggestions are put forward. Ajzen (1991) proposed the theory of planned behavior, which states that the behavioral intention of people is mainly controlled by behavioral attitudes, subjective norms, and perceptual behavior. More specifically, the theory states that the behavior of an individual show positive correlations to social pressure and the willingness of individuals when they conclude that it is easier to carry out a certain behavior. Therefore, the present study assumed that five factors influence the recycling behavior of farmers according to observed data for the recycling of pesticide packaging waste (table 1).

Research hypothesis and variable selection
(1) Individual characteristics of farmers Eckberg and Blocker (1997) determined that in general, men play a more active role in environmental protection than women. A study by Shi and Liu (2012) determined that older people have a higher perception of pollution compared to younger people. Factors such as education level and sense of belonging affect Table 1. A summary of five factors influencing the willingness of farmers to participate in recycling of pesticide containers.
farmers' behavioral attitudes toward waste disposal (Wan et al 2021). The present study assumed that age and education level have positive effects on recycling intention.
(2) Container size Information contained in the existing literature indicates that pesticide packaging is typically of a small capacity (Wei and Wang 2014). Farmers are more likely to dispose of small capacity packaging since larger packaging containers are easier to recycle. The present paper assumed that larger container sizes are easier to recycle.
(3) Perception of the risk of pesticide packaging among farmers Subjective norms and perceptions of the outside world are positively correlated with farmers' intentions (Aliabadi et al 2020, Vaz et al 2020). Farmers are placing increasing attention on environmental pollution, leading to a higher likelihood of farmers taking measures to protect the environment. This result indicates that the awareness of farmers of' the hazards posed by solid waste will affect their behavioral decisions to a certain extent. Farmers increase the recovery of pesticide packaging when they realize the risk of this packaging waste to the environment and human health.
(4) Status of local facilities The local availability of waste disposal facilities is one important factor affecting the agricultural waste disposal behaviors of farmers (Huang and Tian 2014). The presence of local recycling points may increase the recovery of pesticide packaging waste.
(5) Intensity of publicity awareness Awareness of the harmfulness of pesticide packaging and the advocacy of its recovery can improve the awareness of environmental protection among farmers and effectively increase the efficiency of pesticide packaging collection. Therefore, the present study assumed that the intensity of local publicity activities has a positive effect on recycling.

Date sources
The data used in this study were obtained from questionnaire surveys conducted in rural areas in different regions of China. It mainly includes Shenyang, Tieling, Dalian, Dandong and Liaoyang in Liaoning, Puyang, Anyang, Xuchang, Nanyang and Xinyang in Henan, Qingyang and Baiyin in Gansu, Suzhou, Huzhou, Liuzhou, Liuan, Hefei, Chuzhou, Wuhu and Bengbu in Anhui, Shaoyang, Hengyang and Chenzhou in Hunan, Bijie, Zunyi, Liupanshui and Qianxinan Buyi Miao Autonomous Prefecture in Guizhou, and more than two hundred administrative villages were designed. Figure 1 shows the data distribution of the main regions.
To ensure the quality of the survey, the principle of random sampling was adopted in this study. The information collected in the surveys included the basic circumstances of individual farmers, pesticide use, disposal of pesticide packaging waste, and the setting of recycling points. A total of 473 questionnaires were distributed, and after eliminating invalid questionnaires, a total of 459 valid questionnaires were harvested, with a valid return rate of 97.04%. Of the questionnaires returned, 45.53% were from Liaoning Province, 8.71% from Guizhou Province, 5.01% from Anhui Province, 3.70% from Henan Province, 2.40% from Gansu Province, and 1.31% from Hunan Province. In the process of releasing the survey, it has been clearly informed that the survey content and data filled in will be used in the research and published publicly. Informed consent was obtained from all study participants for the pesticide packaging study.

Empirical model
The dependent variable in the present study was the willingness of farmers to recycle pesticide packaging waste. This variable was regarded as a categorical variable with three possible values: (1) willingness to recycle; (2) unwilling to recycle, and; (3) undecided. Therefore, the disordered multi-classification logistic regression model was applied in the survey. The data were transformed into two binary logistic models with the uncertain group as the reference: In equation (1) and (2), P is the probability of farmers being willing to recycle pesticide packaging, α willing and α unwilling are the constants of the corresponding models, β willing j and β unwilling j are the regression coefficients of the JTH independent variable in the corresponding model, and Xj is the jth independent variable, respectively.

Results and discussion
3.1. Basic characteristics of farm households Through the analysis and summary of the questionnaire data, it was found that the survey respondents were mainly young adults under 45 years old, of whom 62.3% were men. In terms of education level, 59.9% of the farmers only received up to nine years of compulsory education (table 2).  In terms of pesticide packaging specifications, mainly small containers, 83.9% of pesticide packaging volume < 500 ml. Farmers in the surveyed areas have a high awareness of the degree of local hazards, with 94.8% of the subjects acknowledging that pesticide packaging waste can have a large or small impact on the surrounding environment. Moreover, most regions (56.2%) contained no pesticide packaging recycling points. The results also indicated the need to increase the intensity of publicity activities related to local recycling. It is gratifying to note that the surveyed farmers are more willing to recycle pesticide packaging waste, with only 5.9% of farmers saying they are not willing to participate (table 2).
In this study, the majority of farmers in each province were willing to participate in pesticide packaging. Among them, all surveyed farmers in Hunan Province were willing to participate in pesticide packaging activities, and the others were Guizhou Province, Henan Province, Liaoning Province, Gansu Province, and Anhui Province in that order (figure 2).

Multivariate collinearity and model likelihood ratio test
Correlation between variables in the questionnaire may lead to errors in the results of the two binary logistic models. Therefore, there was a need to assess each variable (Chen and He 2018). The tolerance and VIF values of sample data were tested for multicollinearity in the SPSS18.0. software. Table 3 shows the results of the analysis.
The tolerance values of all variables were between 0.1 and 1.0, whereas the VIF values were all less than 10. These results confirmed the absence of multicollinearity among variables and relative stability of the constructed model.
The control group was set as those respondents unwilling to participate in pesticide packaging recycling. The explanatory (independent) variables were set as gender, age, schooling, container size, perceived risk, cognitive understanding, number of recovery points, existing collection mode, and publicity, whereas recycling willingness was set as the response variable. A disordered multi-category logistic regression model was constructed in SPSS18.0, following which the main factors affecting the recovery of pesticide packaging and their contribution were analyzed. Table 4 and 5 show the fitting information and likelihood ratio test, respectively.
The significance level of model fitting was less than 0.05, thereby passing the inspection. This result indicated that the disordered multi-classification logistic model could be used to analyze and explain the willingness of farmers to recover pesticide packaging. The results of the likelihood ratio test showed that the variables gender, age, schooling, container size, perceived risk, cognitive understanding, number of recovery points, existing collection mode, and publicity had significant effects on the willingness of farmers to recycle pesticide packaging. However, the significant of the influence of education level was 0.257. Therefore, this variable had no obvious influence on the willingness of farmers to recycle. The model was optimized by removing education level.

Analysis of test results
As shown in table 6, the results of the disordered multi-classification logistic regression analysis showed that each factor had a difference influence on the willingness of farmers to recycle pesticide packaging.

Model 1 Analysis of the willingness of farmers to recycle pesticide packaging
The present study conducted a comparison of the group of farmers not willing to participate in pesticide packaging recycling with that showing a willingness. The results showed that age, container size, perception of risk, number of recovery points, and existing collection model showed significant (P < 0.05) relationships with the willingness of farmers to participate in pesticide packaging recycling. The likelihood of people under 45 years old of being willing to participate in the recycling of pesticide packaging exceeded that of people over 60 years old by a factor of 5.193, whereas there was no statistically significant difference between the 45-60 years old group and the > 60 years old group. The willingness of farmers using pesticide packaging specifications of < 100 ml, 100-500 ml, and 500-1,000 ml exceeded those of farmers using packaging > 1,000 ml by factors of 4.092, 17.130, and 11.596, respectively. Of the people who believed that pesticide packaging is harmful to the environment, 20.8% were willing to participate in the recycling of pesticide packaging. There were no statistically significant differences between the perception of risk of pesticide packaging to the environment and that of other dangers. Willingness to recycle in places with two recycling points exceeded that in which there were no recycling points by a factor of 4.246. In villages, the likelihood of people recycling pesticide packaging in cash mode was 30.2% of those who recycle in other modes.

Model 2 Analysis of the results of the willingness of farmers to recycle pesticide packaging
An analysis of the uncertainty of the results of the logistic model indicated that container size, perception of risk, number of recovery points, and existing collection model had statistically significant (P < 0.05) relationships with the willingness of farmers to participate in pesticide packaging recycling activities. In general, farmers using pesticide packaging with a volume in the range of 100-500 ml were undecided on their willingness to participate in recycling. Willingness to recycle pesticide packaging of less than 1,000 ml exceeded that for packaging with a volume > 1,000 ml by a factor of 11.231. Of the respondents, 24.3% believed that pesticide packaging poses a risk to the environment. The willingness of undecided villagers with no access to local recycling points to participate in recycling exceeded that with access to two recycling points by a factor of 3.915. Respondents in villages indicated that those recycling pesticide packaging in barter and cash payback modes were 7.3% and 23.1% less likely to recycle pesticide packaging in other modes.
From an economic point of view, conducting pesticide packaging waste recycling is a public activity that requires broad participation of farmers (Lu et al 2022). At the same time recycling activities will be regulated by the recycling model, risk awareness, social norms, social trust and other aspects together, is the result of multifactor optimization (Xu et al 2021).
The size of pesticide containers in this paper significantly affects the willingness to recycle, although farmers using small sizes (<500 ml) have a higher willingness to recycle, but given the wide range of small package sizes (Lin 2022), it will directly increase the negative impact of subsequent disposal or resourcefulness. The number of recycling points that already exist and the mode of recycling can also have a significant impact on the willingness to recycle. The willingness to recycle reached 100% in Hunan Province (figure 2), which may be related to its establishment of a recycling and disposal pilot since 2017, and the establishment of a more unified policy through a dedicated system, a commissioned recycling system, and a conscious recycling system (Lv 2020). This provides a practical and effective case for establishing a uniform national recycling policy.
Although this study identified the intervention effect of external environment on pesticide packaging waste, there are still some limitations that deserve further exploration. Since pesticide packaging activities have not yet been replicated nationwide, the small sample area of the study may affect the accuracy of the policy. Therefore, the study area needs to be further expanded and the sample size increased in future studies to improve the accuracy of policy implementation.

Conclusions
The present study used the logistic model to study the influence of various factors on the willingness of farmers to recycle pesticide packaging waste. A survey study collected information from 459 randomly selected respondents from rural areas in six provinces and cities. The present study aimed to explore whether farmers are willing to participate in the recycling of pesticide plastic containers. The research results showed that the individual characteristics of farmers and the intensity of awareness had no influence on the willingness of farmers to recycle. More specifically, gender, educational qualifications, and awareness had no significant influence on whether farmers were willing to participate in the recycling of pesticide packaging waste. In contrast, age had a slight influence on the willingness of farmers to recycle in model 1. Container size, perception of risk, and availability of local recycling facilities also had significant effects on the influence of farmers to participate. Specifically, a container size between 100 ml and 500 ml showed the strongest correlation with willingness to recycle. Most pollution by pesticide packaging was for relatively small containers not conducive to recycling. Therefore, the present study proposes a lower pesticide packaging limit of~500 ml. Awareness among farmers of the potential harm caused by pesticide packaging increased their willingness to participate in recycling. Farmers without access to local pesticide packaging recycling points were more willing to recycle. In addition, farmers were less interested in barter and cash payback modes. Therefore, there is a need to engage with local farmers when establishing recycling points, an appropriate recycling mode, and the management mechanism.
The present study proposes the following countermeasures for the disposal of pesticide packaging waste based on the analysis results: (1) Increasing investment funds for research into degradable pesticide packaging and encouraging farmers to use this packaging.
(2) Increasing efforts to provide constant awareness to farmers of the risks posed by pesticide waste packaging. (3) Encouraging pesticide producers and management enterprises to establish small recycling points that can be conveniently used by farmers. (4) Public or private institutions should play leading roles establishing a complete recycling system, identifying the responsibility of recycling, and facilitating the transportation, storage, and treatment of pesticide container waste.