Training of Farmers’ Cooperatives, Value Perception and Members’ Willingness of Green Production

Abstract

The issue of environmental pollution caused by traditional agricultural production operations is becoming increasingly serious. Farmers are the direct actors of production, and their willingness to green production deserves the greatest attention. Technical training conducted by farmers’ cooperatives worldwide in recent years appears to have changed farmers’ willingness to adopt green production technologies, but there is a lack of empirical testing of the impact mechanisms. Therefore, based on a sample of 1147 members of China’s citrus production cooperatives, we theoretically and empirically explored the impact of this; the mechanism of the effect was analyzed through the endogeneity treatment and robustness test of farmers’ value perception, as well as the instrumental variable method (IV-Oprobit). The results showed that farmers’ overall willingness to adopt green production technologies was low, and increasing the number of training sessions in farmers cooperatives could significantly enhance their willingness. Specifically, the probability of members being “very willing” to adopt technologies increased by 3.2% for each additional training session in cooperatives. Additionally, cooperative training can significantly improve members’ technology applicability and benefit–cost perceptions of green production technologies, thus enhancing their willingness to adopt; both types of value perceptions are important transmission mediators of this effect, and the mediation effects account for 5.98 and 14.53% of the total effect, respectively. Other than that, the results of the heterogeneity analysis showed that the effect of cooperative training on the willingness to adopt them was positively significant regardless of small-, medium- or large-scale members, with the most significant effect on small-scale farmers. This study provides a better understanding of the impact of technical training of cooperatives on farmers’ willingness, contributes to the enrichment of value perception theory, and provides a basis for formulating relevant policies to encourage cooperatives to perform their training function and promote green production in agriculture.

1. Introduction

At present, China is implementing the rural revitalization strategy, and green agricultural development is an important part of it [1], but the average amount of chemical pesticides applied per unit area in China is up to three times the levels seen in developed countries [2,3], far exceeding the international average level [4]. One of the main culprits that aggravates China’s agricultural non-point source pollution and causes the degradation of agricultural ecosystems is the non-green production method caused by the excessive application of chemical pesticides and fertilizers [5]. Non-green agricultural production methods lead to a series of adverse consequences for the ecological environment, such as the use of highly toxic chemical pesticides, excessive pesticide concentration, excessive use of chemical fertilizers, and random disposal of agricultural wastes, etc., [6,7]. Although these behaviors can meet the needs of eliminating diseases and insect pests and crop nutrition, the large amount of chemical residues produced have caused different degrees of pollution to the soil, atmosphere and water environment [8], and even threatened human health [9]. Therefore, the Chinese government has continuously emphasized the promotion of green agricultural development [10] and the harmonious coexistence of man and nature [11]. The problem in response is that agricultural pollution has become one of the most serious environmental pollution problems [12]; as the ancient Chinese saying goes: The fire has burned to the eyebrows, and the green transformation of traditional agriculture is imminent. Green production in agriculture is a green development mode with low input, low consumption and sustainable development of resources. The use of various green production technologies in the early, middle and late stages of production has greatly reduced the impact of agricultural development on the ecological environment. The green agricultural production technology mainly includes the application of chemical pesticide fertilizer, biological pesticide and organic fertilizer, physical pest control, ecological integrated control and so on. Against the backdrop of global advocacy for ecologically sustainable development, agricultural green production technologies have gradually come to the center of the world stage [13]. These mainly include reducing the use of chemical fertilizers and pesticides, increasing the use of biological pesticides and organic fertilizers and other rational field management models. [14]. Green production technologies can not only improve the efficiency of agricultural resource allocation [15], but also reduce environmental pollution and the waste of resources [16], which are important ways to achieve green transformation and development of agriculture [17]. For example, green production technology can reduce the impact on non-targeted organisms and land and water sources, and improve environmental benefits. It can promote fertile land, optimize the quality of agricultural products, and improve economic benefits. It can reduce the input of pesticides and fertilizers, and reduce the input cost of traditional pesticides and other production factors. However, in developing countries, the proportion of farmers’ adopting these technologies is still very low, which is a topic that needs to be studied and solved urgently [18]. As the micro-subject of agricultural production, farmers’ behaviors are completely driven, dominated and directly affected by their will. [19], therefore, the key to green agriculture lies in the greening of farmers’ willingness to adopt green production technologies [20].

The academic community has carried out relatively abundant research on farmers’ willingness to green production. It is generally believed that farmers’ willingness and behavioral results are largely derived from internal endowments and external influences. On the one hand, internal factors are mainly related to resource endowments such as the farmers’ age [21], education level [22], geographic location [23], cognitive ability [24], and family size [25]. From the perspective of external influence, many scholars have found that external factors such as information acquisition [25], technological environment [26], government policy guidance [27], and external risks [28] also play a growing role in affecting farmers’ willingness. In recent years, the most striking example of external influencing factors is the role that agricultural technology training can play in promoting farmers’ adoption willingness [29,30], which can improve farmers’ own access to relevant technologies [31], as well as through the spillover effect to enhance the willingness of the remaining farmers to adopt technologies [32]. In particular, Huang et al. found through empirical research that after farmers participated in the technical training of rice, their understanding and practical ability of green production technology was significantly improved, and the use of chemical fertilizers was continuously reduced [33]. After comparing farmers’ access to new technologies, Yu et al. found that technical training was the most effective method [34]. However, due to the small farmers’ economic awareness [35], the low level of organization [36], and the opportunistic behavior of non-cooperation [37], who will organize and implement effective green production technologies training and education has become a difficult issue. Among the organizations that provide training to farmers, peasant cooperatives embedded in rural social networks can fully demonstrate the advantages of agricultural technology extension by providing face-to-face, repetitive and field training [38]. Unlike American Agricultural Cooperatives, which are small in number and large in size, Chinese farmers’ co-operatives are very large in number, but each co-op is small in size and membership. Farmers’ cooperatives here are generally established according to the type of agriculture, such as vegetables, fruits and so on, which is very conducive to the cooperatives to carry out professional training. They are usually rooted in their villages, and the members of the cooperative are all local farmers who know each other well, which also determines their unique social networks and organizational mobilization capabilities. As mentioned by the International Co-operative Alliance (ICA), the ICA established the Manchester Principles in 1995: co-operatives are organized to develop “education, training and information” [39]. It is undeniable that cooperatives have an irreplaceable training function [40], and they can carry out in-depth technology training in rural societies by virtue of their geographical and organizational advantages. Therefore, as a training organization, farmers’ cooperatives are very developed in developing countries [41], providing a good training approach for agricultural green production technologies [42]. At present, the total number of registered farmers’ cooperatives in China has reached 2.217 million, and the number of farmers who have joined has exceeded 100 million, including more than 50% of farmers. Many studies have verified that cooperatives have played a positive role in enhancing farmers’ willingness to adopt green production technologies [43,44,45], and have provided us with useful lessons. However, few studies have analyzed the effect of training provided by cooperatives in promoting farmers’ willingness to green production through empirical research and conducted heterogeneity analysis on farmers with different endowments, and there is also a lack of research to explain the transmission mechanism of this effect.

Therefore, the main purpose of this study was to conduct a more rigorous effect and mechanism assessment of the impact of the technical training provided by cooperatives on farmers’ willingness to adopt them, so as to give full play to the role of cooperatives in technology promotion. This paper is conducive to the expansion of research related to farmer cooperatives and green production. First, we empirically analyzed the impact of cooperative technical training on farmers’ adopt willingness. Second, we used instrumental variables to control for endogeneity issues arising from the bidirectional causal relationship between technical training participation and intentions to adopt technologies. Third, since farmers with different planting scales show huge differences in receiving training and choosing green production technologies, we conducted a heterogeneity analysis on the differential impact of farmers’ planting scales. Fourth, we analyzed the mediating role of farmers’ value perception in this influence relationship. Since previous studies have emphasized the role of perception in technology adoption, but the connotation and extension of value perception have not been able to obtain a unified opinion, this study further draws on the achievements in the field of technology adoption and deconstructs perceived value into technology applicability perception and benefit–cost perception, as well as explores farmers’ value perception of green production technologies and its mediation effect.

2. Theoretical Analysis and Research Hypothesis

2.1. Theoretical Analysis

The human capital theory founded by Schultz and Becker pointed out that education and training are an important way for organizations to improve individual human capital, which helps to mobilize the enthusiasm and initiative of individuals to learn independently and accumulate in practice, and effectively improve human resources. As a cooperative economic organization voluntarily established by farmers, its training advantages are: first, the trainers of cooperatives and farmers have mutual understanding and familiarity with each other [46], and individual farmers have more trust in the cooperative trainers [47] and the organization [48] and are more likely to adopt technologies after receiving their training. Second, cooperatives can accurately determine which type of green production technology is suitable for the production needs of local farmers, and consider the individual needs of farmers [49] to improve the efficiency of agricultural technology extension. Third, cooperatives are all established in local villages, which have prominent geographical advantages [50], and can bring convenient information access for trained farmers [51]. Fourth, peasant cooperatives are established by farmers spontaneously, and the training using the local language system is more in line with their cognitive dimensions [38], and is more conducive to farmers to absorb the training information [47]. Fifth, the cooperatives usually provide farmers with pre-production, production and post-production supporting services after training to help them reduce the cost of adopting green production technologies, and farmers are more likely to adopt relevant technologies after training [52]. Technical training for members in cooperatives can not only strengthen farmers’ technical thinking and modernization awareness, and effectively stimulate their willingness and ability to adopt new technologies [38,53], but also promote the overall development of human capital in cooperatives. According to this analysis, the cooperatives’ training and education can promote farmers’ accumulation of professional human capital in green production technologies, and to a certain extent, enhance the awareness and ability of members to independently carry out agricultural green production.

For farmers, training is an effective way to improve the value perception of new things [54]. Recent studies have shown that, for farmers, training is an effective way to improve the perception of new things [55]. Training can enhance farmers’ value perception of green production technologies [56], and this perception has a significant difference in their behavioral willingness [57], It is not difficult to judge by this derivation that farmers’ value perception may be the key factor affecting willingness to adopt green production technologies [58]. The core of value perception is defined as “the trade-off between perceived benefit and effort”, which is generally deconstructed into perceived benefit and perceived effort. Premkumar and Bhattacherjee compared the technology acceptance model, the expectation confirmation model and their mixed models and found that the mixed model has the best explanatory power for individual behavioral intentions [59].

Thong et al. used a technology acceptance model that integrated perception of technical applicability and compared it with the original model and found that the model integrated with perceived ease of use had the strongest explanatory power [60]. Numerous studies have confirmed that the user’s perception of usefulness and applicability and the perceived level of benefit and cost can directly affect the adoption behavior of the product [61,62,63,64].

Value perception theory believes that people’s willingness to a product, service or behavior is based on the individual’s sense of value and experience of them, and is a subjective comprehensive evaluation formed by weighing and comparison [65]. Perception comes from the individual’s processing of the acquired information, that is, through the individual’s cognitive logic of things, the results before and after the behavior are simulated, compared and judged with their own expectations. When the perceived profit is greater than the perceived loss, the greater the individual’s value perception level, the more obvious their willingness is, and the greater the probability of the implementation of the behavior. [66]. When combined with the existing research, this paper deconstructs it into technology applicability perception and benefit–cost perception [60,61,62,63], in which benefit–cost perception can be further divided into economic benefit, ecological benefit and cost input perception. By participating in cooperative training, members may perceive economic, ecological and environmental benefits, input costs and the application of green production technologies, leading to an an improved perception level of green production technologies value; this willingness to promote green production is triggered after the brain’s comprehensive judgment on the benefits, which is consistent with the existing research results [67]. Under different perception levels, the resource allocation equilibrium point is different, and their behavioral intentions are different [68], and farmers’ value perception of green production technologies determines their behavior choices. After the technical training of farmers’ cooperatives, the majority of members have been promoted to understand green production technology, and they can perceive the potential benefits [69], and enhance the subjective initiative of adopting technology from the subjective level of individuals.

The cost–benefit theory proposes that, as rational economic persons, farmers will make rapid and correct production decision-making adjustments according to changes in economic returns, and maximize returns by reconfiguring production factors [70]. The theory of the “ecological economic man” believes that the people in the ecological economic system not only have the economic rationality to pursue the maximization of economic benefits, but also have the ecological rationality to attach importance to the value of the ecological environment [71]. Therefore, the decision of members as to whether they adopt green production technologies is usually based on the perception of economy, ecological environment and cost input. If the marginal benefit of an action is greater than the marginal cost, it will increase the willingness to perform the action, otherwise it will not [72]. The higher the perception level of the green production technologies value of members, the deeper the understanding of the overall benefits of technology adoption [73], and the more likely it is to germinate the endogenous motivation to adopt it. A high level of value perception can also reduce the cost of searching and processing relevant information [74], and promote the willingness of members to adopt green production technologies.

2.2. Theoretical Framework

The key to farmers’ willingness to adopt lies in the return benefits. As a rational economic person, their behavioral decisions are based on a comprehensive judgment after rationally weighing and comparing costs and benefits [75]. This provides theoretical support for examining the willingness of farmers from the perspective of value perception theory. According to the previous analysis, the technology applicability perception and benefit–cost perception constitute the value perception, and the perception can be further divided into economic benefit, ecological benefit and cost input perception. Strengthening training can help enhance members’ value perception in green production technologies, and the improvement of their perception level can promote their participation. Logical derivation shows that cooperative training can promote willingness through the mediation effect of members’ value perception. Therefore, a theoretical model of members’ willingness to adopt green production technologies was constructed (see Figure 1).

2.3. Research Hypothesis

According to the theoretical and model analysis, the green production willingness of members will be affected by the training for cooperatives, and the behavioral willingness and decision-making of the individual will be affected by the value perception level of the behavior. The higher the cooperatives’ training level, the higher the individual’s technology application perception and benefit–cost perception of green production, and the more they can perceive the potential value, thereby stimulating them to generate stronger psychological intentions. Members’ value perception of green production technologies actually acts as a mediator in this influence mechanism. Accordingly, this paper proposes three research hypotheses:

Farmer cooperative training has a significant positive impact on members’ willingness to adopt green production technologies, and the technology applicability perception and benefit–cost perception play a mediating role in the effect of cooperative training on the members’ willingness.

3. Materials and Methods

3.1. Data Sources and Sample Characteristics

To control the endogeneity issues caused by the difference of agricultural industries, this paper selected the members of citrus cooperatives with high homogeneity as the research object. The research data comes from household surveys conducted by the research team in August 2020 and August 2021, covering 14 large counties (districts) for late-ripening citrus cultivation in China. The sampling method was to randomly select 2–4 townships in each sampled county, then randomly select 1–4 farmers’ cooperatives in the selected town-ships, and then randomly select 5–10 members from the selected cooperatives as the survey objects. A total of 1147 members from 148 cooperatives were selected for this survey. The contents of the survey include education of cooperatives, members’ cognition, GCT adoption, individual characteristics, and family endowments. After statistics and sorting, sample data can be seen in (

Table 1. Sample size distribution.

City	County	Sample Size	Percentage/%
Chengdu	Pujiang	60	5.23
	Jintang	74	6.45
Nanchong	Gaoping	69	6.02
	Nanbu	53	4.62
	Pengan	80	6.97
Meishan	Dongpo	60	5.23
	Renshou	115	10.03
	Danling	114	9.94
Ziyang	Anyue	71	6.19
	Yanjiang	119	10.37
Neijiang	Zizhong	113	9.85
Yibin	Jiangan	126	10.99
Dazhou	Dachuan	40	3.49
	Quxian	53	4.62

).

3.2. Variables Definition and Description

3.2.1. Explained Variable and Explanatory Variable

According to the relevant research on the setting of variables [11,38], the dependent variable in this paper is the members’ willingness to green production, which is represented by “Are you willing to adopt green production technologies?”, with options ranging from “unwilling” to “very willing”, which were assigned values of 1–5, respectively. In general, the interviewed members’ willingness is at a general level, with an average value of 2.537, indicating that the current farmers’ adoption level of green production technologies is still in its infancy. The explanatory variable is the number of technology trainings of cooperatives, which is represented by “How many times did you participate in green production technologies trainings of cooperatives last year?”. The average number of members participating in the training is 4.184, showing that the farmer cooperative has a good training foundation, which is basically consistent with reality.

3.2.2. Mediator Variable

According to theoretical analysis and relevant research [60,61,62,63], this paper divides members’ value perception of green production technologies into technology applicable perception and benefit–cost perception, and this perception is further divided into economic benefit, ecological benefit and cost input perception. The overall perception level of the green production technology of the sample members is low: only 22.8% of the members believe that the technology is suitable for their own production operations, indicating that the current green production technology still has the characteristics of high threshold, and only a small number of groups can master the technology proficiently. The average value of benefit–cost perception is 1.309. Specifically, 51.3% of members perceive that green production technologies have economic value, 52.5% of members perceive that they have ecological value, and only 27% of members perceive that they will reduce cost input. More than half of the members can perceive the economic and ecological environmental benefits brought by green production technologies, but most people’s perception that they reduce the input cost of chemical pesticides is still at a low level.

3.2.3. Control Variables

Referring to the existing research [21,22,23,24,25], the control variables mainly include two parts: personal characteristics and family endowments. This paper controlled for 12 variables such as gender, age, education level, and the family income of members. The specific meaning and assignment of the variables are shown in

Table 2. The meaning and assignment of variables.

Variables		Variable Definitions	Variable Assignment	Mean	Std. Dev.	Min	Max
Explained variable		willingness to adopt green production technologies	Unwilling = 1; less willing = 2; generally = 3; somewhat willing = 4, very willing = 5	2.537	1.018	1	5
Explanatory variables		Cooperatives’ training	Number of times members participated in cooperative training last year	4.184	3.716	0	50
Mediator variable	Technology applicability perception	Perception that green production technology is applicable to oneself	Applicable = 1, Not applicable = 0	0.228	0.420	0	1
	Benefit-cost perception	Perception of economic value	High economic value = 1, general economic value = 0	0.513	0.500	0	1
		Perception of ecological value	High ecological value = 1, general ecological value = 0	0.525	0.500	0	1
		Perception of cost input	Cost input reduction = 1, cost input flat = 0	0.271	0.445	0	1
		Total perception of benefit-cost	The superposition of economic value, ecological value and cost input perception	1.309	0.874	0	3
Control variables	Personal characteristics	Gender	Male = 1, Female = 0	0.717	0.451	0	1
		Age	Actual age/years old	55.717	9.528	25	83
		Education level	Actual years of education/year	7.494	3.514	0	18
		Political identity	Whether to be a member of the CPC: Yes = 1, No = 0	0.153	0.361	0	1
		Cadre status	Whether a village cadre or civil servant: Yes = 1, No = 0	0.129	0.335	0	1
		Health level	Very low = 1; low = 2; average = 3; high = 4, very high = 5	4.065	0.803	0	5
	Family Endowment	Household income	Last year’s total household income/ten thousands yuan	28.951	89.630	0	1500
		Planting scale	Area of citrus planted by families/mu	26.105	82.572	0	1200
		Planting years	Number of years of citrus cultivation	11.762	8.882	1	50
		Geographical location	Distance of the household from the cooperative/km	1.996	3.033	0	35
		Social network	Whether there are relatives or friends working in the agricultural sector: Yes = 1, No = 0	0.198	0.585	0	15
		Village terrain	Topography of the village: plain = 1; hilly = 2; mountainous = 3	2.039	0.366	0	3
Instrumental variable		Neighbor Influence	There are neighbors around who use excessive amounts of fertilizers and pesticides: a lot = 1; more = 2; average = 3; less = 4, very little = 5	1.511	0.726	1	5

.

3.3. Research Method

3.3.1. Ordered Probit Model

Since the explanatory variables are ordered by variables from 1 to 5, the Oprobit model was used to estimate the effect of cooperative training on members’ willingness to adopt technologies. The empirical model was set up as follows.

$${\mathrm{willingness}}_i={\alpha }_0+{\alpha }_1{\mathrm{training}}_i+{\alpha }_2{X}_i+{\mu }_i$$

(1)

where ${\mathrm{willingness}}_i$ represents the adopted willingness of member; ${\mathrm{training}}_i$ is the green production technology training of the cooperative; $X_i$ is a series of control variables, individual characteristics of members, family characteristics, etc.; ${\mu }_i$ is a random disturbance term. Assuming μ ~ N (0, 1) distribution, the Oprobit model can be expressed as follows.

$$\begin{array}{l}P(\mathrm{willingness}=1|x)=P({\mathrm{willingness}}^{*}\le r_0\left|x\right)=\phi \left(r_0-{\alpha }_1{\mathrm{training}}_{\mathrm{i}}-{\alpha }_2{X}_i\right)\\ P(\mathrm{willingness}=2|x)=P(r_0<{\mathrm{willingness}}^{*}\le r_1\left|x\right)=\phi \left(r_1-{\alpha }_1{\mathrm{training}}_i-{\alpha }_2{X}_i\right)-\phi \left(r_0-{\alpha }_1{\mathrm{training}}_i-{\alpha }_2{X}_i\right)\\ P(\mathrm{willingness}=5|x)=P(r_3\le {\mathrm{willingness}}^{*}\left|x\right)=1-\phi \left(r_3-{\alpha }_1{\mathrm{training}}_i-{\alpha }_2{X}_i\right)\end{array}$$

(2)

In Equation (2), $r_0$<$r_1$<$r_2$<$r_3$ are the parameters to be estimated; ${\mathrm{willingness}}_i$ represented as values 1 to 5, respectively, indicate being “unwilling” to “very willing”. By constructing the likelihood function, the parameters of the model were estimated by using the great likelihood method.

3.3.2. Instrumental Variable Method

In fact, both cooperative training and willingness may have the same impact on certain common reasons, such as individual members’ participation in the cooperative organization. Therefore, this paper uses the instrumental variables approach (IV-Oprobit) to address the issue of biased estimation results due to endogeneity. With the selection conditions based on the fact that instrumental variables should be highly correlated with the endogenous explanatory variables but not with the nuisance terms [76], this paper selects non-green production behaviors such as overuse of fertilizers and pesticides by neighbors as the instrumental variables of the model. There is a strong correlation between a farmer’s level of value perception of something and the perception and acceptance of that something by others in that farmer’s environment [67]. Farmers’ level of value perceptions may be influenced by the perceptions and behaviors associated with their neighbors in the same village, but the neighbors’ willingness to act is not directly related to the study participants’ own willingness to do. Then, two regression models were constructed to test them, and the results showed that the effect of the instrumental variable on members’ willingness was not significant, but the effect on cooperative training was significant at the 1% level, and both correlation tests passed, again indicating that the selection of this instrumental variable was reasonable.

3.3.3. Mediation Effect Model

To further verify the mediation effect of perceived value between cooperatives’ training and willingness to green production, referring to the mediating effect test proposed by MacKinnon et al. [77], the model was set as follows.

$${\mathrm{willingness}}_i={\alpha }_0+{\alpha }_1{\mathrm{training}}_i+{\alpha }_2{X}_i+{ϵ}_i$$

(3)

$${\mathrm{cognitivelevel}}_i={\beta }_0+{\beta }_1{\mathrm{training}}_i+{\alpha }_2{X}_i+{\mu }_i$$

(4)

$${\mathrm{willingness}}_i={\gamma }_0+{\gamma }_1{\mathrm{training}}_i+{\gamma }_2{\mathrm{cognitivelevel}}_i+{\alpha }_2{X}_i+{\phi }_i$$

(5)

where, ${\alpha }_1$ in Equation (3) reflects the total effect of cooperatives’ training on members’ willingness to adopt technologies, ${\beta }_1$ in Equation (4) indicates the effect of cooperatives’ training on the mediator variable value perception of green production technologies, and ${\gamma }_1, {\gamma }_2$ in Equation (5) indicate the direct effects of both cooperatives’ training and green production value perception on the i-th member’s willingness to adopt them, respectively. Substituting Equation (4) into Equation (5) yields the mediating effect ${\beta }_1{\gamma }_2$, and the ratio of the mediation effect to the total effect ${\beta }_1{\gamma }_2/{\alpha }_1$ reflects the size of the mediation effect.

4. Results and Discussion

4.1. The Impact of Farmer Cooperatives’ Training on Members’ Willingness to Adopt Green Production Technologies

Model (1) in

Table 3. The impact of farmer cooperatives’ training on members’ willingness to adopt green production technologies.

Variables	Model (1) Oprobit	Model (2) IV-Oprobit	Model (3) Marginal Effect/%
			Unwilling	Less Willing	Generally	Somewhat Willing	Very Willing
Cooperatives’ training	0.049 *** (0.009)	0.234 *** (0.037)	−0.071 *** (0.015)	−0.005 (0.005)	0.017 *** (0.004)	0.027 *** (0.004)	0.032 *** (0.017)
Control variables
Gender	−0.212 *** (0.072)	−0.140 ** (0.063)	0.042 ** (0.017)	0.003 (0.005)	−0.010 (0.008)	−0.016 (0.011)	−0.019 *** (0.007)
Age	0.003 (0.004)	0.002 (0.003)	−0.001 (0.001)	0.000 (0.000)	0.000 (0.000)	0.000 (0.000)	0.000 (0.000)
Education level	0.045 *** (0.012)	0.028 ** (0.012)	−0.008 *** (0.003)	−0.001 (0.001)	0.002 (0.001)	0.003 (0.002)	0.004 *** (0.001)
Political identity	0.337 *** (0.101)	0.223 ** (0.092)	−0.067 *** (0.025)	−0.005 (0.007)	0.016 (0.012)	0.026 (0.017)	0.030 *** (0.010)
Cadre status	0.084 (0.107)	0.062 (0.073)	−0.019 (0.022)	−0.001 (0.003)	0.005 (0.006)	0.007 (0.009)	0.008 (0.010)
Health level	0.145 *** (0.042)	0.088 ** (0.039)	−0.027 ** (0.011)	−0.002 (0.003)	0.006 (0.005)	0.010 (0.007)	0.012 *** (0.004)
Household income	0.000 (0.000)	0.000 (0.000)	0.000 (0.000)	0.000 (0.000)	0.000 (0.000)	0.000 (0.000)	0.000 (0.000)
Planting scale	0.001 * (0.000)	0.001 * (0.000)	0.000 * (0.000)	0.000 (0.000)	0.000 (0.000)	0.000 (0.000)	0.000 * (0.000)
Planting years	0.000 (0.004)	−0.001 (0.002)	0.000 (0.001)	0.000 (0.000)	0.000 (0.000)	0.000 (0.000)	0.000 (0.000)
Geographical location	0.013 (0.011)	0.012 (0.007)	−0.003 (0.002)	0.000 (0.000)	0.001 (0.001)	0.001 (0.001)	0.002 (0.001)
Social network	0.078 (0.054)	0.048 (0.039)	−0.015 (0.011)	−0.001 (0.002)	0.004 (0.004)	0.006 (0.005)	0.007 (0.005)
Village terrain	−0.020 (0.089)	−0.021 (0.059)	0.006 (0.018)	0.000 (0.001)	−0.002 (0.004)	−0.002 (0.007)	−0.003 (0.008)
Pseudo R2/lnsig_2	0.045	1.309 *** (0.021)
LRchi2/atanhrho_12	145.20 (0.000)	−0.971 *** (0.383)
Log likelihood/waldchi2	−1533.814	761.96 (0.000)

Notes: ***, **, and * denote significance levels of 1%, 5%, and 10%, respectively; the numbers in parentheses are robust standard errors of the coefficients; all data in the table are rounded. lnsig_2 is the significance test value of the second-stage estimating equation and atanhrho_12 is the error term correlation test of the first- and second-order estimating equations. The following table is the same.

examines the direct effect of farmer cooperatives’ training on members’ willingness to adopt them, and the results show that, controlling for variables such as gender and age, cooperative training has a significant positive effect on members’ willingness to adopt them, and is significant at the 1% level, which indicates that the more frequent cooperative training is, the more likely members have the willingness. After the members participate in the training of farmer cooperatives, they can get more useful information about green production technology through the communication of the trainer in easy-to-understand rural language. Through the rural social network built by the cooperatives, farmers can also consult their neighbors who also attended the training, as well as directly ask the cooperative’s professional and technical staff to deepen their mastery of technology and the accumulation of specialized human capital. This result is consistent with the findings of the previous theoretical derivation.

Model (2) uses IV-Oprobit to deal with the endogeneity issue, and the estimation shows that the coefficient of cooperative training is significantly positive, again testing the positive effect of farmer cooperatives’ training on members’ willingness to adopt green production technologies. The model lnsig_2 value is 1.309 and the two-stage estimation of the model is significant and also passed the likelihood ratio test with atanhrho_12 test, proving that the cmp method is superior to Oprobit estimation in this model and the use of instrumental variable is valid. Model (3) presents the results of the marginal effect estimation. From the results, the effect of cooperatives’ training on members’ willingness is similar in direction and significance to the baseline regression results reported in Model (1) after controlling for endogeneity issues, again validating this positive effect. The results showed that for every one unit increase in cooperative training, the probability of members being “very willing” to adopt them increased by 3.2%.

In terms of other control variables, endowments such as gender, age, and education level had significant effects on willingness to adopt them to varying degrees and were significant at the 1 to 10% level. Among them, female members were more willing to participate in green production and were 1.9% more likely to be very willing to participate than males, probably because females were more concerned about the potential impact of ecological environment on their family’s health. A possible explanation for the education level is that the more years of education a member has, the better his or her understanding and cognitive ability [47], the more motivated to learn new things, and the more likely they are to obtain information about green production technologies from the cooperative training, improve value perceptions, and thus develop a willingness to adopt them. CCP members are capable representatives of rural areas who possess big-picture awareness and are more receptive to advanced technologies, pay more attention to rural ecological civilization issues, and are usually able to exert a demonstration-driven effect in the choice of green production technologies [67]. Green production relies on more manpower to get rid of pests and requires more labor, and the better the individuals’ health, the more energy and stamina they can provide to the community members. In addition, the planting scale positively influenced the members’ willingness at the 10% level, and the larger the household’s citrus planting area was, the more likely they were to obtain higher economic benefits through green production technologies [78].

4.2. Heterogeneity Analysis of Members’ Planting Scale

Members with different resource endowments differ greatly in their values and preferences, and different values promote various cognitive and behavioral styles [25,30,67]. The farmers’ household cultivation scale is closely related to their behavioral intentions [79], and the land cultivation scale has a significant effect on farmers’ willingness and behavior of green production technologies [80], which was similarly verified by the above empirical results. Therefore, this paper combined the actual situation of Chinese farmers’ citrus planting scale, using 5 and 20 mu planting areas as the dividing line, dividing small, medium and large scale members, and conducting Oprobit and IV-Oprobit regressions on the effects of the green production intentions of the three groups of members, respectively. As shown in

Table 4. Effect of cooperative training on the willingness of different generations of members’ willingness to adopt green production technologies.

Variables	Small Scale Group (≤5 mu)		Medium Scale Group (5–20 mu)		Large Scale Group (≥20 mu)
	Oprobit	IV-Oprobit	Oprobit	IV-Oprobit	Oprobit	IV-Oprobit
Cooperatives’ training	0.045 *** (0.015)	0.284 *** (0.056)	0.040 *** (0.014)	0.227 *** (0.087)	0.064 ** (0.020)	0.173 ** (0.077)
Control variables	Controlled	Controlled	Controlled	Controlled	Controlled	Controlled
Sample size	514	514	434	434	199	199
Virtual R2	0.043	—	0.033	—	0.044	—
Waldchi2/LRchi2	59.01 (0.000)	508.69 (0.000)	40.19 (0.000)	270.49 (0.000)	25.56 (0.000)	48.38 (0.000)
Loglikelihood	−651.991	−1975.837	−589.320	−1795.930	−275.402	−835.358

Notes: *** and ** denote significance levels of 1% and 5% , respectively.

, after controlling for the endogeneity issue, cooperative training had a significant positive effect on members’ willingness to adopt them in all small, medium and large groups of members, and all three groups of members were able to better accumulate individual specialized human capital for green production technologies through cooperative training, which in turn enhanced their willingness. However, there are some differences among the three groups, in which cooperative training positively affected the green production intentions of small and medium-sized group members at the 1% significance level, and positively acted on large-sized group members only at the 5% level. This may be due to the fact that members with larger citrus production scales have resource endowments with more favorable conditions, and the large producers themselves are mostly local technicians and professionals who can obtain less useful information from farmers’ cooperative training, and thus are less affected by the cooperative training than small-scale farmers. On the other hand, small and medium-scale farmers, limited by their own endowments, generally rely more on spontaneously established farmers’ cooperatives, trust them more, and are not only willing to take the initiative to participate in the training, but also are willing to accept the beneficial content of technical training at the subjective level, and thus are more influenced by the cooperatives’ training.

4.3. Mediation Effect Test

In order to corroborate the theoretical analysis that the perceived value of green production technologies plays the role of a mediator in this influence, this paper provides an in-depth analysis of the mediating transmission mechanism through which cooperative training finally influences the willingness of the members by changing the technology applicability perception and benefit–cost perception, and tests hypotheses with stepwise regression, the results of which are presented in

Table 5. Mediation effect of technology applicability perception.

Variables	Model (4) Technology Applicability Perception (Oprobit)	Model (5) Willingness to Adopt Green Production Technologies (Oprobit)	Model (6) Willingness to Adopt Green Production Technologies (IV-Oprobit)
Cooperatives’ training	0.040 *** (0.012)	—	0.201 *** (0.056)
Technology applicability perception	—	1.535 *** (0.089)	1.162 *** (0.090)
Control variables	Controlled	Controlled	Controlled
Sample size	1147	1147	1147
LRchi2/Waldchi2	126.91 (0.000)	429.99 (0.000)	1048.32 (0.000)
Loglikelihood	−552.899	−1391.417	−1381.704
lnsig_2	—	—	1.309 *** (0.021)
atanhrho_12	—	—	−0.750 ** (0.393)

Notes: *** and ** denote significance levels of 1% and 5% , respectively.

and

Table 6. Mediation effect of benefit-cost perception.

Variables	Model (7) Benefit-Cost Perception (Oprobit)	Model (8) Willingness to Adopt Green Production Technologies (Oprobit)	Model (9) Willingness to Adopt Green Production Technologies (IV-Oprobit)
Cooperatives’ training	0.041 *** (0.009)	—	0.182 *** (0.062)
Benefit-cost perception	—	0.510 *** (0.039)	0.389 ** (0.101)
Control variables	Controlled	Controlled	Controlled
Sample size	1147	1147	1147
LRchi2/Waldchi2	88.63 (0.000)	252.14 (0.000)	666.76 (0.000)
Loglikelihood	−1379.147	−1480.131	−4604.783
lnsig_2	—	—	1.309 *** (0.021)
atanhrho_12	—	—	−0.725 ** (0.392)

Notes: *** and ** denote significance levels of 1% and 5% , respectively.

. Model (4) in Table 5 is the regression analysis conducted between cooperative training and the technology application perception by the members, and it was found that cooperative training significantly contributed to the perception with a coefficient of 0.040 and passed the significance test at the 1% level. Model (5) was used to examine the effect of technology application perception on the willingness to promote green production, and the results showed that by removing the cooperative training variable, there was also a significant positive effect of technology application perception on the willingness at the 1% level; this suggests that the higher the level of members’ technology application perception, the more likely they were to have willingness. Model (6) introduces independent and mediator variables, and the estimated coefficients of both cooperative training and technology applicability perception are positive; the effects on willingness are both significant at 1% level. Further analysis shows that the marginal effect of cooperative training on members’ willingness is 0.201, which is lower than the corresponding marginal effect of cooperative training when members’ perceptions are not introduced, 0.234, showing that the technology applicability perception plays a partial mediating effect in this. The Sobel (Sobel method) and the self-sampling (Bootstrap method) were used to conduct the test, and the results showed that the mediation effect for the members’ technology applicability perception was significant at the 1% level, its size was about 0.014, and the share in the total effect was about 5.98%, and the above results confirmed the robustness of the mediation effect. The results indicate that the technology applicability perception plays a partially mediating role in the process of the cooperative training influencing members’ willingness to adopt them. Specifically, with the deepening of the cooperative’s training, the members’ manipulation and proficiency of green production technology has gradually increased, and their technology applicability perception level has also improved, and thus the probability of being willing has been enhanced accordingly.

The regression analysis of cooperatives’ training and members’ benefit–cost perception of green production technologies using Model (7) in Table 6 revealed that cooperatives’ training significantly contributed to members’ perceptions with a coefficient of 0.041 and passed the significance test at the 1% level. Model (8) was used to examine the effect of members’ benefit–cost perception on their willingness to adopt green technologies, and the results showed that by removing the cooperatives’ training variable, there was also a significant positive effect of the perception on willingness at the 1% level; this indicates that the higher the level of benefit–cost perception, the more likely members were to have willingness to adopt the technologies. Model (9) introduces independent and mediator variables, the estimated coefficients of both cooperatives’ training and benefit–cost perception are positive, and the effects of both on the willingness to adopt green production technologies are significant at the 1% level. Further analysis shows that the marginal effect of cooperatives’ training on members’ willingness is 0.182, which is lower than the corresponding marginal effect of cooperatives’ training when members’ perceptions are not introduced, 0.234, showing that benefit–cost perception plays a partial mediation effect in this. Similarly, the Sobel test (Sobel method) and the self-sampling test (Bootstrap method) were used for the regressions, and the results showed that the mediation effect of benefit–cost perception was significant at the 1% level, and its size was about 0.034, which accounted for about 14.53% of the total effect; the above results confirmed the robustness of this mediation effect. The results indicate that the benefit–cost perception plays a partially mediating role in the process of cooperatives’ training influencing members’ willingness to green production. Specifically, as the cooperatives have deepened their training, the members experienced more of the increased economic and ecological benefits and reduced costs brought by green production technologies, the overall level of benefit–cost perception increased, and thus the probability of being willing increased, proving that the hypothesis was valid.

4.4. Robustness Test

To further ensure the reliability of the estimation results, this paper performed a robustness test on the main effects in two aspects, as detailed in

Table 7. Robustness test results.

Variables	(10) Oprobit	(11) IV-Oprobit	(12) Probit	(13) IV-Probit
Cooperatives’ training	0.050 *** (0.009)	0.232 *** (0.038)	0.044 *** (0.012)	0.258 ** (0.041)
Control variables	Controlled	Controlled	Controlled	Controlled
Virtual R2	0.046	—	0.100	—
LRchi2/waldchi2	146.17 (0.000)	721.59 (0.000)	103.92 (0.000)	606.32 (0.000)
Loglikelihood/Logpseudolikelihood	−1531.282	−4651.582	−469.370	−3563.805

Notes: *** and ** denote significance levels of 1% and 5% , respectively.

.

On the one hand, the sample robustness test is performed. Because the exploration was about the effect of farmer cooperatives’ training on members’ willingness, the farmers aged over 80 years were less able to receive training and understand information, and their current willingness to produce was weakly associated with cooperative training. Therefore, the Oprobit and IV-Probit models were used for estimation after removing the sample of older members, as detailed in Models (10) and (11), and the results were still significant at the 1% significance level, indicating good sample robustness.

On the other hand, model robustness tests are performed. The members’ willingness was classified as a dichotomous variable, and those who were “generally”, “unwilling” and “less willing” were classified as the “no willingness” sample group, while those who were “more willing” and “very willing” were classified as the “willingness” sample group, and then estimated using the Probit and IV-Probit models, as detailed in Models (12) and (13); the results showed that the regression results were consistent with the previous paper, indicating that the robustness of the model estimates was good.

5. Discussion

5.1. Similarities and Differences with Existing Research

As you can see, our results are consistent with those reported by Wu [81], with calculations by Luo et al. [67], showing that technical training provided by agricultural cooperative can significantly increase farmers’ propensity to adopt green production techniques. In the citation and validation of the human capital theory, our conclusions, like those of Zhang et al. [82] and Khan et al. [83], confirm that technical training can increase the capital accumulation of technical specialization of farmers. In terms of heterogeneity analysis, we found the same findings as Chen et al. [80], that the impact on the willingness of small-and medium-scale farmers to adopt green production technologies was stronger on the basis of confirmation of the usefulness of training. In the aspect of value perception, this paper reconfirms the empirical results of Liu et al. [38]. Farmers’ perception of the economic value of green production technology is an important intermediary variable that influences their adoption of technology, however, compared with the results of Wang et al. [64], this paper draws a different conclusion, that is, as an important component of benefit–cost perception, farmers’ perception of the environmental benefits of green production technology plays a part in it. The possible reason is that, according to the theory of the “Eco-economic man”, farmers are the most direct contacts in the eco-economic system; at the same time, there will also be ecological rationality that attaches importance to the value of the ecological environment [71]. The greater the awareness of farmers of the ecological benefits that can be obtained by adopting green production technologies, the more likely they are to make judgments on the improvement of the comprehensive benefits of adopting technologies, and then choose to adopt these technologies.

5.2. Theoretical Implications

Based on the above theoretical analysis and empirical estimation, three implications were obtained from this study. In terms of theoretical implications, first, the study findings further validate and affirm the human capital theory, pointing out that technology training can indeed promote farmers’ human capital specialization in green production technologies and facilitate their acceptance of new technologies. Second, the research findings further enrich the value perception theory, especially with the help of research results from other technology adoption fields, and have highlighted a discussion around the connotation and extension of value perception. Third, the application and innovation of benefit–cost theory in the field of green production is enriched by including the costs and benefits of production technologies in the analytical framework for theoretical exploration.

5.3. Limitation and Future Research Direction

A limitation of this study is that we only focused on the current status of farmers’ willingness to adopt green production technologies in the sample, but ignored farmers’ persistent willingness, which is more likely to motivate people to eventually implement this behavior. If we can introduce the experimental economics method and conduct multiple surveys and data collection targeting the sample farmers, and finally form a panel database of farmers’ willingness to sustain adoption, we can draw more convincing conclusions through empirical tests. Meanwhile, in recent years, it should be noted that the technical training carried out by agricultural enterprises may also have a good promotion effect, which may be a direction for our future research. Furthermore, given the increasing importance of technical training provided by farmer cooperatives in increasing farmers’ willingness to green production, there is a need to further include studies from other crops and other regions to test the external validity of our findings.

6. Conclusions and Recommendations

The problem of environmental pollution such as fertilizer and pesticide abuse caused by traditional agricultural production operations is becoming increasingly serious, especially in developing countries. “The fire has burned to the eyebrows”, the transformation of traditional agriculture to green production is imminent, and farmers are the direct actors of production; their willingness to adopt green production technologies deserves the greatest attention. However, farmers’ willingness is generally low due to the high threshold of green production technology and their own low level of learning ability and organization. As farmer cooperatives are deeply embedded in rural society, training conducted by such organizations seems to change farmers’ behavioral intentions. To test this conjecture, this study empirically explored the impact of technical training provided by cooperatives on farmers’ willingness and its influence mechanism from the perspective of value perception, using a sample of Chinese citrus cooperative members.

6.1. Conclusions

Based on 1147 micro-survey data of members in Sichuan province, we explored the influence mechanism of farmer cooperatives’ training on members’ willingness to adopt green production technologies and analyze the mediating role of value perception in it. The Oprobit model and mediation effect model were used for analysis, and the instrumental variables approach (IV-Oprobit) was used for the endogeneity treatment and robustness test. First, the overall farmers’ willingness is low, and increasing the number of training for farmers’ cooperatives can significantly improve the willingness of members to adopt them. Specifically, for each additional training session in the cooperative, the probability of members being “very willing” to participate in adoption increased by 3.2%. Second, cooperatives’ training can significantly enhance the members’ value perceptions of green production technologies in terms of technology applicability and benefit–cost, i.e., both value perceptions are important mediators of this effect, and their mediation effects account for 5.98 and 14.53% of the total effect, respectively. Third, the results of the heterogeneity analysis of the land planting scale showed that the effect of cooperatives’ training on the farmer’s willingness was positively significant regardless of small-, medium- or large-scale members, with the most significant effect on small-scale farmers.

6.2. Policy Recommendations

Based on the above empirical estimation, this study proposes three realistic policy recommendations. First, cooperatives’ training has a significant driving effect on members’ willingness to adopt green production technologies. Accordingly, a sound training and education system for green production technologies in farmers’ cooperatives should be established, and relevant training should be carried out by giving full play to the organizational advantages of cooperatives in order to drive members to participate in it. Second, cooperative training can improve members’ willingness through the mediating role of value perception, and improving farmers’ value perception level can help enhance their endogenous motivation to adopt green production technologies. Therefore, it is necessary to strengthen the economic benefits, ecological benefits, and cost input explanation of green production technology in order to enhance the technical application perception and benefit-cost perception for members. Third, we should design differentiated training and education programs for farmer groups with different land cultivation scales.