Whether the Use of the Internet Can Assist Farmers in Selecting Biopesticides or Not: A Study Based on Evidence from the Largest Rice-Producing Province in China

Abstract

The pivotal measure for reducing pollution and facilitating green and sustainable agriculture lies in the application of biopesticides to replace chemical pesticides. The argument still rests on whether the use of the Internet can assist farmers in selecting biopesticides or not. In light of 532 microscopic research datapoints from the largest rice-producing province in China, the Probit model was applied in this article to probe the influencing factors regarding the use of the Internet on their selection of the biopesticides by farmers, and the TAM-PR model was also adopted to explore its intrinsic mechanisms. According to the research findings, the use of the Internet directly contributed to the application of biopesticides, which can affect farmers’ decisions regarding biopesticides by means of perceived usefulness and perceived ease of use in an indirect manner, and the mediating effect was 19.74% and 20.98%, respectively. The result regarding perceived risk was not significant. The use of the Internet has a remarkable effect on farmers with high incomes and large-scale operations, while it has an insignificant effect on farmers with low incomes and individual operations. It has a remarkable positive effect on farmers with high academic qualifications compared with those with low academic qualifications. Furthermore, personal, household-based production and village and green cognitive characteristics also have a significant influence on the application of biopesticides. Hence, it is of great significance to continuously facilitate the application of rural Internet usage, encourage environmentally friendly modes of production, and reduce agricultural pollution.

1. Introduction

The extensive use of pesticides in agricultural production leads to a reduction in soil beneficial bacteria, air pollution, and serious harm to human health [1,2]. Therefore, the use of biopesticides to replace traditional pesticides is an important measure to promote the green and sustainable development of agriculture and ensure the food safety of producers and consumers. Since 2015, China has focused on the green development of agriculture, implemented drug reduction actions, and vigorously developed biopesticide technology. At present, the proportion of large-scale biopesticide enterprises with an annual output value of more than RMB 20 million has risen to one-seventh, and the approved single-dose varieties of biopesticides have reached 20% of the total number of pesticides. The authorized patents rank third in the world, with an annual output value of more than RMB 1.85 billion [3], and the development of biopesticide technology has become a long-term strategic goal in China. However, strangely, Chinese farmers do not buy biopesticides, and even then, there are differences in consciousness and behavior. The main reasons why farmers do not use biopesticides are a lack of awareness, difficulty in changing application habits, doubts about the application effect, and the use of chemical pesticides by people in surrounding areas [4]. Therefore, how to effectively match the demand for biopesticide informatization with the demand for agricultural production incomes, narrow the distance between farmers’ opinions, and let biopesticides enter the homes of ordinary farmers has become an important topic in the promotion of biopesticide technology. The rapid development of the Internet in rural areas is conducive to solving this problem. The informational advantages of the Internet are sufficient to meet farmers’ needs for biopesticide information [5] and effectively match farmers’ production and drug needs through the advantages of comprehensive searches, accurate positioning, convenience, and speed and encourage farmers to choose agricultural production methods with more comparative advantages. During the thirteenth Five-Year Plan period, the number of rural Internet users in China reached 309 million, and the rural Internet penetration rate rose to 55.9% [6]. The Internet has been integrated into all fields of farmers’ production processes and lives, subtly changing farmers’ traditional production habits and integrating into the many links of agricultural production decision-making, providing new production elements and feasible paths for farmers to choose environmentally friendly production methods such as biopesticides. Therefore, this article delves into the adoption behavior of farmers for biopesticides from the perspective of Internet use, and tries to answer the following questions: Does the use of the Internet promote farmers’ choices of biopesticides? What is the relationship between farmers’ Internet use and biopesticide selection? Are there differences between farmers with different resource endowments? Through the discussion of the above issues, it is of great significance to reveal the implications of the use of the Internet for farmers to choose biopesticides, which is of great significance for promoting green agricultural production technology and improving the rural ecological environment.

Scholars at home and abroad have not reached a consensus on the application of biopesticide technology in farmers. Since the implementation of the Action Plan for Zero Growth in Pesticide Use by 2020 [7], the use of biopesticides has increased year by year. As an environmentally friendly green pesticide, biopesticides provide new ideas for reducing the number of pesticides. However, due to the lack of green production awareness and high costs, biopesticides have not been rapidly popularized, and most farmers remain on the sidelines of the adoption of biopesticide technology [8]. Guo et al. (2021) found that rice farmers are more likely to apply biopesticides because biopesticides are less prone to drug resistance and have basically no toxic side effects, which is conducive to improving the quality of agricultural products [9]. Huang et al. (2022) believed that the effective promotion of biopesticides requires farmers to see long-term benefits so as to drive surrounding farmers to adopt biopesticides, which induces herd behavior and deviations in the willingness of farmers to accept biopesticides [10]. In addition, scholars have conducted research on the factors influencing farmers’ use of biopesticides on rice, including internal factors such as risk attitude [11], age, education level, land size [12], the purpose of growing food, food flavor [13], contamination by external factors such as the environment [14], participation in training programs [15], government regulation [16], market intervention [17], etc. In contrast, research analyzing the variability and limitations of farmers’ adoption of biopesticides from the perspective of Internet use has not yet made a breakthrough.

Changes in farmers’ behavioral decision-making through Internet use have attracted the attention of many scholars. The main reason why the Internet can gradually change their decision-making behavior is that farmers have Internet thinking. Using the Internet, farmers expand their information needs in agricultural production, encourage other farmers to expand cultivated land, improve the allocation of labor resources, increase production decision-making and production efficiency, and increase peasant incomes [18]. Researchers have found that the ease of access the Internet provides to knowledge and information influences farmers’ adoption of green production technologies [19]. Chen et al. (2022) found that the use of the Internet significantly increased the adoption rate of straw return technology and enhanced the effectiveness of land conservation [20]. The Internet has had a profound impact on farmers’ adoption of green production technologies. However, some scholars believe that farmers need agricultural production information and are not interested in learning new technologies [21], and it is difficult to achieve the purpose of learning Internet technology before obtaining production information. Moreover, ordinary smallholder farmers rely more on past experiences to make agricultural production decisions [4], and the transition to the use of the Internet is still a long process.

The purpose of this article was to examine whether the use of the Internet would help farmers adopt biopesticides and how the use of the Internet influences farmers’ adoption of biopesticides. Specifically, there are three main contributions in this article: First, existing studies have been conducted from the perspective of the influencing factors of pesticide use by farmers, while fewer studies have studied farmers’ production behavior decisions from the Internet perspective. Nowadays, Internet access is being improved in the rural areas of developing countries, and the Internet is undoubtedly an important factor influencing farmers’ adoption of biopesticides. This study constructs a theoretical framework of Internet use and farmers’ adoption of biopesticides and further reveals the role of the relationship between the two. Second, the TAM-PR model was constructed to explore the intrinsic influence mechanism of the use of the Internet and farmers’ adoption of biopesticides. Internet use affects farmers’ decisions on whether to adopt biopesticides through perceived usefulness, perceived ease of use, and perceived risk. Reasonable explanations are provided for farmers’ decisions on whether to adopt biopesticides. Finally, the heterogeneity of the use of the Internet by farmers to adopt biopesticides was analyzed. The variability of the use of the Internet by farmers to adopt biopesticides was explored in terms of farmers’ education, economic level, and scale of operation. The findings can help to promote targeted biopesticide expansions in rural areas of China.

2. Theoretical Analysis and Research Hypothesis

2.1. The Influence of the Internet on the Decisions of Farmers to Adopt Biopesticide Technology Based on the TAM Model

Davis’s theory of technology acceptance is mainly used to analyze how external variables influence users’ acceptance behavior regarding new technologies by perceived usefulness and ease of use. The use of the Internet has enabled farmers to better determine the benefits of agricultural production information. Therefore, the use of Internet applications as external variables to directly affect the perceived usefulness and perceived ease of use helps promote farmers’ attitudes and willingness to choose biopesticides.

Perceived usefulness refers to the economic and ecological benefits of farmers’ use of biopesticides. As an effective channel for information integration, the Internet can effectively break cognitive barriers, broaden the information channels of production materials, grasp the benefits of green production, and urge farmers to choose green and ecological agricultural production methods based on the concept of sustainable development so as to maximize the benefits of agricultural production. The Internet can show farmers the environmental improvement of demonstration villages that have adopted biopesticide technology through videos and live streaming. This will stimulate farmers’ yearning for a better environment and strengthen their determination to improve the rural ecological environment by adopting biopesticides. Furthermore, the Internet can also display the short-, medium- and long-term costs, benefits, and ecological impacts of biopesticides replacing chemical pesticides through charts, text editing, and other functions so that farmers can abandon the quantitative comparison of economic benefits and ecological benefits, see the long-term benefits of biopesticides and the short-term benefits of excessively using chemical pesticides, and enhance farmers’ understanding of the economic benefits and ecological impacts of pesticide reduction and applications. This can encourage farmers’ adoption of biopesticides by helping them understand the economic and ecological benefits of reducing pesticide use. Perceived ease of use means that the easier it is for farmers to perceive the technology or method when adopting a biopesticide, the stronger the willingness to use it. Due to the transcendental and acquaintance trust mechanisms of traditional agricultural production [22], there are two problems with farmers’ perceived ease of use: Firstly, agricultural production relies on kinship and geography, resulting in farmers’ aversion to the use of biopesticides [23]. Secondly, the information asymmetry between traditional agricultural knowledge and modern agricultural knowledge leads farmers to believe that the use of biopesticides is more complicated [24]. The advantage of the Internet breaks geographical restrictions, enhances farmers’ understanding of biopesticides, strengthens farmers’ grasp of knowledge details, and makes the actual operation easier. Using WeChat, QQ (Tencent Instant Messenger), APP (Application) to achieve timely communication with farmers can ensure that farmers accurately grasp the operation process of biopesticides, ensure that farmers fully understand the application principle of biopesticides, and improve the operability of farmers so as to promote the use of biopesticides instead of chemical pesticides under their own initiative [25]. Based on this, this article proposes the following hypotheses:

Hypothesis 1:

The use of the Internet can increase farmers’ perceptions of the usefulness of biopesticides and encourage them to adopt biopesticides.

Hypothesis 2:

The use of the Internet can improve farmers’ perceptions of the ease of use of biopesticides and encourage them to adopt biopesticides.

2.2. The Effect of Internet Use on the Decisions of Farmers to Adopt Biopesticide Technology Based on the PR Model

Perceived risk (PR) stems from elements of the theory of perceived value and is often used to analyze the uncertainty that individuals perceive due to information asymmetry. Chinese farmers often have difficulty recognizing and accepting new things because of information asymmetry. The emergence of the Internet has helped farmers break through the barriers of information transmission, allowing farmers to experience the simplicity and convenience and the benefits of new technologies. As a rational economic person, the ultimate goal of a farmer in agricultural production is to maximize production efficiency, and reducing the use of chemical pesticides will inevitably cause farmers to worry about yield and quality. Communication facilities are the best means to update agricultural production knowledge and solve the problems of subjective production experience judgment [26]. First of all, by relying on the Internet’s live broadcasts, short videos, and other channels, farmers can be taught in detail about the application of biopesticides, and the obstacles of insufficient and unskilled learning of farmers can be eliminated. The second is to disseminate biopesticide knowledge through text, pictures, videos, and other forms, which is convenient for retention and convenient for farmers to learn anytime, anywhere. Cost-free repetitive learning can help farmers correctly understand all aspects of biopesticides and sort out and correct misunderstandings [27]. On the one hand, this can enhance farmers’ awareness of the safety of biopesticides, break the stubborn problems of empirical production modes of the past, and correctly help them understand the feasibility of biopesticides. On the other hand, it enhances farmers’ awareness of biopesticide standardization, stimulates farmers’ technical ability, stimulates farmers’ interest in trying, helps eliminate farmers’ risk concerns, and boldly encourages them to try biopesticides. Based on this, this article proposes the following hypothesis:

Hypothesis 3:

The use of the Internet can reduce farmers’ perceptions of the risk of biopesticides and encourage them to adopt biopesticides.

2.3. TAM-PR Theory Integration Analysis

This study integrates the TAM and PR models, analyzes the influencing factors of Internet use on farmers’ use of biopesticides, and forms an extended theoretical model of farmers’ adoption of biopesticides (Figure 1), which has stronger explanatory power for the TAM-PR integration model. Using the three-dimensional paths of the economic and ecological benefit perceptions of usefulness, perceived operation comprehensibility, and new technology production safety and standardization perceptions, the decision-making mechanism of Internet use regarding farmers’ biopesticide applications is discussed. Based on this, this article proposes the following hypothesis:

Hypothesis 4:

Internet use can encourage farmers to adopt biopesticides.

3. Materials and Methods

3.1. Sample Selection

Heilongjiang Province is the most important rice-producing region in China. Its rice sowing area and yield accounted for 12.87% and 13.67% of the national total, respectively [28], ranking it first in the country. The selection of farmers in Heilongjiang Province as research objects is typical and representative. The main reasons are as follows: First, Heilongjiang Province, as the most important grain-producing area in China, has many farms and state-owned enterprises, and large-scale operations are suitable for the promotion of biopesticide technology. Second, the province has a number of biopesticide enterprises and scientific research institutes, a leading biopesticide technology level, and sufficient production capacity, which makes it feasible to promote this technology. Third, many places in the province have been selected as national agricultural and rural informatization demonstration bases, indicating that Heilongjiang Province’s rural Internet construction is good and the Internet penetration rate is high. The data used in this article are taken from a field investigation conducted by our research group in Heilongjiang Province, China, from January 2022 to March 2022 with the theme of “biopesticide applications in rice farmers”. The survey team first selected 11 typical rice-producing counties and then used a combination of stratified sampling and simple random sampling to randomly select 15 to 20 households in 3 villages in each county. The questionnaire conducted “one-on-one” interviews with key agricultural operators. The questionnaires selected by the survey team mainly related to the basic situation of farmers’ households, production and operations, village infrastructure, awareness of green production, and the use of biopesticides. In the end, the survey obtained 532 valid samples.

From the basic characteristics of the respondents, males predominate, accounting for 57.14%. In total, 58.83% of the respondents were aged 35 to 55, 9.21% were under 35 years old, and 31.96% were over 55 years old. The education level of the respondents was generally not high, and 51.88% of the respondents had an education level below junior high school. According to the area of cultivated land operated by households, the average area of cultivated land plus circulating land is 257.4 mu. Furthermore, the income from cultivated land was 97.74%, less than RMB 50,000. In general, respondents are basically in line with the views of Chinese rural households with low incomes, lower education levels, and higher degrees of aging (

Table 1. Distribution of sample.

Variable	Definition	Sample Size	Proportion (%)
Gender	Female	228	42.86
	Male	304	57.14
Age	≤35	49	9.21
	35–55	313	58.83
	≥55	170	31.96
Land size	100–200 mu	136	25.56
	200–300 mu	212	39.85
	≥300 mu	184	34.59
Education level	Never attended school	78	14.66
	Primary school	198	37.22
	Junior high school	203	38.16
	Senior high school or above	53	9.96
Farming incomes	≤20,000 RMB	166	31.20
	20,000–50,000 RMB	354	66.54
	≥50,000 RMB	12	2.26

Note: Respondents with more than three years of experience in growing rice.

).

3.2. Methods

Since a farmer’s decision to adopt biopesticides is a binary discrete variable, a binary Probit model with strong explanatory power for behavioral decision-making should be selected to study the influencing factors of farmers’ use of biopesticides in the context of the Internet. The basic model is:

$$P\left(Y_i=1 | I_i\right)=\phi \left({\alpha }_i+{\beta }_i+{\gamma }_i{X}_i+{\epsilon }_i\right)$$

(1)

In Formula (1), $Y_i$ represents the variable explained in this section, namely, the farmer’s decision to adopt a biopesticide. The explanatory variable for biopesticide technology is “1” and “0” if not used. $I_i$ represents the Internet usage by respondents; $X_i$ is a vector indicator of the control variable; and ${\alpha }_i$, ${\beta }_i$, and ${\gamma }_i$ are the estimated coefficients of the regression model. ${\epsilon }_i$ is a random error term.

3.3. Variables and Measurements

3.3.1. Core Explanatory Variable

Studies have pointed out that the use of the Internet determines the use of terminal devices, such as mobile phones and computers, by farmers to obtain information. Therefore, this article is designed around the following question: “Do you use the Internet to get information about agricultural production?” Farmers using the Internet to obtain information are defined as 1; otherwise, 0. In terms of Internet use, 329 households (61.84%) in the sample area used the Internet to search for agricultural production information, and 203 households (38.16%) did not use the Internet to search for agricultural production information. This indicates that respondents are more enthusiastic about using the Internet to search for agricultural production information and that there is greater potential for learning about agricultural technology through the Internet.

3.3.2. Explained Variable

The explained variable is biopesticide adoption behavior. Scholars measured the adoption of two biopesticides using dummy variables: pesticide application or amount [29] and household pesticide application behavior [30]. To avoid difficulties in measuring the size of pesticide spray containers, dummy variables are used to measure biopesticide adoption. The questionnaire question “Are you using biopesticides instead of chemical pesticides in your agricultural production this year?” was used to make measurements. The value is 1 if it is adopted by the farmer, and otherwise, it is 0. In addition, in order to better determine whether farmers use biopesticides, farmers are asked, “Do you know the biopesticide varieties (Bacillus subtilis, B. thuringiensis, etc.?” to test the reliability of the questionnaire data.

3.3.3. Control Variables

For the purposes of avoiding the interference of other factors, this article selects control variables based on five aspects: personal characteristics, farmers’ cognitive characteristics, household production and operation characteristics, village characteristics, and green cognitive characteristics [31,32]. Personal characteristics include gender, age, health status, and education. The characteristics of household production and operation include the amount of cultivated land, agricultural labor, agricultural technical training, agricultural incomes, etc. The characteristics of the village include the ease of access to the village, the distance from the village to the township, and the village infrastructure. Farmers’ green cognitive characteristics include pesticide packaging handling, health awareness, and environmental awareness.

3.3.4. Mediators

Perceived usefulness refers to the economic and ecological benefits of farmers’ adoption of biopesticides [33]. Perceived ease of use means that farmers can easily understand and use biopesticides. Perceived risk refers to a farmer’s ability to use biopesticides in a safe and regulated manner. Therefore, the perceived usefulness, perceived ease of use, and perceived risk are measured using a Likert scale. We use a five-point Likert scale with 1 point for “highly disagree”, 2 for “somewhat disagree”, 3 for “neutral”, 4 for “somewhat agree”, and 5 for “highly agree”, and a secondary scale (three questions) is designed to measure the results. The arithmetic mean is then used to measure the mean, which measures the perceived usefulness, perceived ease of use, and perceived risk.

3.3.5. Instrumental Variable

Farmers’ behavioral decisions are based on their subjective behavior, which can be missing variables and uncontrollable factors, resulting in a bias in variable measurements. It is necessary to discuss endogenous issues. Tool variables are an effective way to mitigate endogenous problems. One should choose tool variables that are highly correlated with the explanatory variables without directly affecting the explanatory variables. Therefore, this article chooses “How often do you go online each week?” and “How many mobile phones do you have?” as instrumental variables; the main reason is that the frequency of farmers’ weekly Internet access can reflect the time farmers spend online and the time they spend searching for information, which has a direct impact on Internet use but does not directly affect farmers’ adoption of biopesticides. Specific indicators are shown in

Table 2. Variable assignment and descriptive statistics.

Variables	Variable Definition	Mean			T Test
		Total N = 532	Unused N = 203	Used N = 329
Explanatory variable
Internet use	Do you use the Internet to get information about agricultural production? No = 0, Yes = 1	0.618	0	1	-
Explained variable
Biopesticides adoption	Are you using biopesticides instead of chemical pesticides in your agricultural production this year? No = 0, Yes = 1	0.412	0.350	0.450	0.100 ***
Control variables
Personal characteristics
Gender	Female = 0, Male = 1	0.571	0.552	0.583	0.031 *
Age	Respondent’s actual age	50.60	50.95	50.39	0.56
Health status	What is the physical health of the respondents? Very healthy = 1, More healthy = 2, Average = 3, Unhealthy = 4, Very unhealthy = 5	2.259	2.246	2.267	−0.021
Education	What is your level of education? Never went to school = 1, Primary School = 2, Junior middle school = 3, High School and above = 4	3.434	3.438	3.432	0.006 **
Family production and operation characteristics
Land size	What is the arable land area of your household?	257.4	255.7	258.4	−2.7
Agricultural technology training	Have you ever participated in agricultural technology training? No = 0, Yes = 1	0.622	0.603	0.634	0.031 *
Agricultural labor force	How many people in your family work mainly in agriculture?	1.641	1.601	1.666	0.065 *
Farming incomes	How much is your family’s total annual incomes from growing rice?	2.978	2.925	3.010	0.085
Village characteristics
Village transportation	How accessible are the villages? Very inconvenient = 1, Inconvenient = 2, Generally = 3, Convenient = 4, Very convenient = 5	2.562	2.310	2.717	0.407 **
Distance from village to township	How many kilometers is the distance from your village to the township?	2.49	2.13	2.72	0.59
Village infrastructure	What about the village infrastructure? Very poor = 1, Poor = 2, Generally = 3, good = 4, Very good = 5	2.077	2.133	2.043	−0.090 *
Green cognitive characteristics
Packaging treatment	Do you deal with pesticide (or biopesticide) packaging? No = 0, Yes = 1	0.440	0.438	0.441	0.003
Health awareness	Do you think biopesticides are harmless to humans? Rarely = 1, Less = 2, Generally = 3, More = 4, Very much = 5	2.408	2.453	2.380	−0.073 **
Environmental awareness	Do you think biopesticides can protect the environment? Rarely = 1, Less = 2, Generally = 3, More = 4, Very much = 5	2.258	2.197	2.295	0.098
Mediators
Perceived usefulness (1~5)	Biopesticides have an environmental protection effect.	2.177	2.123	2.210	0.087 ***
	Biopesticides can let me get better profits.
	Biopesticides help me to reduce the use of chemical pesticides.
Perceived ease of use (1~5)	Learning about biopesticide technology is very easy.	2.466	2.419	2.495	0.076 ***
	Through simple training, biopesticide technology can be easily mastered.
	Through technical explanation, it is easy to understand the basic principles of biopesticide.
Perceived risk (1~5)	Worried that applying biopesticides will not effectively remove pests and diseases.	2.532	2.567	2.511	−0.056
	Worried that biopesticides and chemical pesticides cannot be applied together.
	Worried about not skillfully mastering biopesticides, resulting in poor results.
Instrumental variable
Frequency of Internet use	How often do you go online every week? Seldom = 1, Sometimes = 2, Often = 3, Usually = 4, Always = 5	2.229	2.157	2.274	0.117 ***
Number of mobile phones	How many mobile phones do you have?	1.366	1.132	1.510	0.378 ***

Note: *, **, and *** represent significance at the statistical levels of 10%, 5%, and 1%, respectively (same below). Differences were compared using a parametric t-test. Perceived usefulness, ease of use, and risk, from 1 to 5, respectively, represent: highly disagree, somewhat disagree, neutral, somewhat agree, and highly agree.

for descriptive statistics.

4. Results

4.1. Baseline Regression

For the purposes of preventing the multicollinearity problem between variables, the VIF test was carried out in this article, and the results showed that there was no serious multicollinearity problem between the variables (VIF < 2). As shown in

Table 3. Estimates of Internet use on interviewees’ adoption of biopesticides.

Variables	Model 1		Model 2		Model 3
	Coefficient	SE	Coefficient	SE	Coefficient	SE
Explanatory variable
Internet use	0.2613 ***	(0.1142)	0.2634 ***	(0.1157)	0.2681 ***	(0.1134)
Control variables
Personal characteristics
Gender			0.3527 ***	(0.1109)	0.3487 ***	(0.1148)
Age			−0.0071 *	(0.0073)	−0.0074	(0.0078)
Health status			−0.0792	(0.0503)	−0.0731	(0.0515)
Education			0.1164 *	(0.0642)	0.1179 *	(0.0653)
Family production and operation characteristics
Land size			−0.0002	(0.0003)	−0.0002	(0.0003)
Agricultural technology training			0.1857 ***	(0.1108)	0.1764 ***	(0.1162)
Agricultural labor force			0.2079 *	(0.1137)	0.1766	(0.1247)
Farming incomes			0.1712 *	(0.1852)	0.1358	(0.1979)
Village characteristics
Village transportation			0.0293	(0.0521)	0.0260	(0.0568)
Distance from village to township			−0.0632	(0.0398)	−0.0754 *	(0.0416)
Village infrastructure			0.0048 *	(0.0034)	0.0052 *	(0.0036)
Green cognitive characteristics
Packaging treatment			0.2096 *	(0.1113)	0.2473 **	(0.1159)
Health awareness			−0.0098 **	(0.0542)	−0.0092 **	(0.0561)
Environmental awareness			0.0442	(0.0561)	0.0446	(0.0593)
cons	2.628 ***	(0.0754)	−1.323 **	(0.5246)	−0.6828	(0.6340)
Regional control	No		No		Yes
Obs.	532		532		532
LR chi2	71.46 ***		58.25 ***		73.61 ***

Note: *, **, and *** represent significance at the statistical levels of 10%, 5%, and 1%, respectively (same below).

, the chi-square value of Model 1 to Model 3 is significant at 1%, indicating that the overall model is highly significant. Model 1 is the regression result adding only the Internet application variable, and the results show that Internet use can significantly promote the adoption of biopesticides. Model 2 to Model 3 were estimated after adding control variables, but Model 3 controlled for regional effects. From the results, it can be seen that the use of the Internet can still significantly encourage the adoption of biopesticides by farmers, which verifies hypothesis 4. From the regression results, it can be seen that the marginal efficiency of Model 2 is smaller than that of Model 1 and Model 3, and the fitting of Model 3 is better. Therefore, the analysis results of the latter option Model 3 are more objective and reliable. Since the implementation of the China Zero Growth Action Plan for Pesticide Use by 2020, biopesticides, as green-friendly pesticides, have always been the best substitute for chemical pesticides. However, due to farmers’ resistance to new agricultural production methods, it is difficult to find ways to have biopesticides enter the homes of ordinary farmers; the popularization of the Internet in rural areas has directly broken the blockade of geographical restrictions, knowledge restrictions, and operational restrictions, allowing farmers to have the opportunity to understand the cost, benefits, and operation methods of biopesticide, and provide them with technical and psychological support to promote the adoption of biopesticides. In addition, in the context of Internet use, factors such as gender, education level, agricultural technology training, agricultural labor force, urban distance, village infrastructure, pesticide packaging and handling, and health awareness all have a significant effect on farmers’ adoption of agricultural technology and biopesticides.

4.2. Endogeneity and Robustness Test

Consider that the model may also have setting errors caused by missing variables, sample selection, etc. Since the IV Probit model can only solve cases where endogenous variables are continuous, “Do you use the Internet to get information about agricultural production?” was selected in this article as a binary dummy variable; hence, it is the conditional mixing procedure (CMP). The relationship between Internet use and farmers’ adoption of biopesticides was re-estimated using the instrumental variable “Internet frequency of use”. As shown in Table 3, Model 4 and Model 5 are the result of the CMP estimate. The first phase showed that the frequency of Internet use had a significant positive effect on Internet use, demonstrating that Internet use, as an instrumental variable, satisfies the relevance requirement. In addition, further testing the exogeneity of Internet usage frequency, the athrho and atanhrho_12 of the model are significant at the statistical level of 1%, indicating that the use of the CMP model is reasonable, and the estimation results are objective and accurate. The second stage shows that the application of the Internet has a significant positive impact on the adoption of biopesticides at a statistical level of 1%, indicating that the use of the Internet can improve the probability of farmers adopting biopesticides, and endogenous testing is reasonable.

Explaining variable substitution, core variable retesting, sample restriction, and other methods are used to further ensure the robustness of estimation results. As shown in

Table 4. Regression results of endogeneity discussion and robustness test.

Variables	CMP		Model 6	Model 7	Model 8	Model 9
	Stage I Model 4	Stage II Model 5	Replace Explanatory Variable	Replace Explanatory Variable	Male Samples	Female Samples
Internet use		1.794 *** (0.2289)	0.4077 *** (0.0692)		0.3784 * (0.1317)	0.5203 *** (0.1523)
Frequency of Internet use	0.0197 * (0.0093)
Number of mobile phones				0.2137 ** (0.0982)
rho_12	−0.7354 *** (0.1695)
atanhrho_12	−1.085 *** (0.4208)
Control variables	Yes	Yes	Yes	Yes	Yes	Yes
Regional control	Yes	Yes	Yes	Yes	Yes	Yes
Obs.	532	532	532	532	342	238

Note: *, **, and *** represent significance at the statistical levels of 10%, 5%, and 1%, respectively (same below).

, in Model 6, the re-estimation of the explaining variables for farmers’ willingness to use biopesticides shows that the application of the Internet significantly promotes farmers’ willingness to use biopesticides, which is consistent with the baseline regression. Taking mobile phone ownership as the explanatory variable to re-estimate Model 7, the results show that mobile phone ownership has a significant positive effect on the use of biopesticides, and the coefficient and direction are small, leading to a baseline regression with more stable results. Model 8 and Model 9 use the restricted sample method, divide the samples into male farmers and female farmers, and re-estimate the impact of Internet use on farmers’ biopesticide use, and the results show that the effect of the coefficient of Internet use is significant and positive for male and female samples on farmers’ biopesticide use, which is consistent with the baseline regression and further clarifies that the results showing that Internet use actively encourages farmers’ to use biopesticides are objective and robust.

4.3. Heterogeneity

Different resource endowments of farmers also have different impacts on the use of biopesticides. Farmers need a certain knowledge reserve to learn how to use the Internet, and large farms with good economic conditions and a high degree of scale are more likely to accept the promotion of new technologies. Therefore, this article selects income, education, and scale management level for heterogeneity analysis. According to the research and collation of data, a net income of RMB 20,000 and below is a low-income group, and a higher than RMB 20,000 income is a high-income group. As shown in

Table 5. Regression results for heterogeneity test.

Variables	Model 10	Model 11	Model 12	Model 13	Model 14	Model 15
	Low Income	High Income	Low Education	High Education	Individual Operation	Large-Scale Operation
Internet use	0.1782 (0.1508)	0.3975 *** (0.1236)	0.1917 * (0.1392)	0.3764 *** (0.1491)	0.1886 (0.1539)	0.2715 ** (0.1304)
Control variables	Yes	Yes	Yes	Yes	Yes	Yes
Regional control	Yes	Yes	Yes	Yes	Yes	Yes
Obs.	367	165	283	249	237	295
LR chi2	43.13 ***	57.24 ***	60.12 ***	53.71 ***	56.94 ***	72.56 ***

Note: *, **, and *** represent significance at the statistical levels of 10%, 5%, and 1%, respectively.

and Model 10 and Model 11, being in a high-income group significantly encouraged farmers to adopt biopesticide technology, while low-income group is not encouraged; it can be seen that it is easier to obtain effective agricultural production information with better economic conditions, while low-income farmers have blocked information channels, strong dependence on chemical pesticides, and weak use of biopesticides.

From the perspective of education level, this article takes whether farmers have completed compulsory education as the dividing line. Since China’s compulsory education is a nine-year system, the number of years of education below nine years is the low-education group, and the higher education group is for more than nine years. As shown in Model 12 and Model 13, the regression coefficient is more significant in the highly educated group than in the low-educated group, indicating that, with respect to the Internet as a technological advancement, different levels of education lead to different outcomes regarding its use [34]. Families with higher education levels are often better at using the Internet to understand biopesticides, while farmers with lower education levels cannot make better use of the Internet, resulting in slower learning about biopesticides [35]. From the perspective of the scale of operations, it is divided into two categories: large-scale operations and individual operations. As shown in Model 14 and Model 15, being in the large-scale operation group significantly encourages the use of biopesticides by farmers, while uptake among the individual operation group has not been significantly improved, indicating that farmers’ adoption of biopesticides may be limited. Farmers with large-scale operations have higher incomes, lower production costs, a strong risk-bearing ability, stronger acceptance of new technologies and new things, and more willingness to learn about biopesticides. They pursue excess returns through technological upgrades. However, farmers with individual operations have a lower risk tolerance and prefer traditional agricultural operations, even if the application of the Internet makes it less difficult to adopt biopesticides.

4.4. Further Mechanism Testing

For the purpose of further analyzing the transmission mechanism of biopesticides caused by Internet use among farmers, the following model was constructed according to the operation suggestions provided by Baumeister et al. [36].

$$Y=cX+{\epsilon }_1$$

(2)

$$M=aX+{\epsilon }_2$$

(3)

$$Y=c^{\prime }I+b{X}_n+{\epsilon }_3$$

(4)

As shown in

Table 6. Impact mechanism of applying the Internet to the adoption of biopesticide technology by farmers.

Variables	Model 16	Model 17	Model 18	Model 19	Model 20	Model 21	Model 22
	Main Regression	Perceived Usefulness	Behavioral Decision	Perceived Ease of Use	Behavioral Decision	Perceived Risk	Behavioral Decision
Internet use	0.2681 *** (0.1134)	0.2123 *** (0.0908)	0.2497 *** (0.0932)	0.2019 * (0.0977)	0.2433 *** (0.0929)	0.1032 (0.0977)	0.2552 *** (0.0928)
Perceived usefulness			0.2439 *** (0.0913)
Perceived ease of use					0.2786 ** (0.0982)
Perceived risk							0.2517 ** (0.1005)
Control variables	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Regional control	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Obs.	532	532	532	532	532	532	532
LR chi2	73.61 ***	55.79 ***	70.24 ***	62.77 **	71.32 ***	49.91	72.68 ***

Note: *, **, and *** represent significance at the statistical levels of 10%, 5%, and 1%, respectively.

, Model 16 and Model 17 show that the application of the Internet has facilitated the adoption of biopesticides and that the use of the Internet can promote the perceived usefulness of biopesticide adoption among farmers. Model 18 shows that perceived usefulness can significantly encourage farmers to adopt biopesticides. A further comparison of Model 16 and Model 18 showed that the influence coefficient of Internet use on biopesticide use decreased from 0.2681 to 0.2497, indicating that perceived usefulness has a partial mediating effect, of which the indirect mediating effect is 19.74%, and hypothesis 1 is verified. Models 19 and 20 show that Internet use can promote the perceived ease of use of biopesticides, and perceived ease of use can significantly promote households’ adoption of biopesticides. Compared with Model 16 and Model 20, the influence coefficient decreased from 0.2681 to 0.2422, indicating that the perceived ease of use has a partial mediation effect, of which the indirect influence mediation effect is 20.98%, and hypothesis 2 is verified. According to Model 21, the results of Internet use on the perceived risk of biopesticides are not significant, indicating that the perceived risk does not have an intermediary effect, and hypothesis 3 is been verified. However, Model 22 suggests that perceived risk can significantly promote farmers’ adoption of biopesticides. The possible reason for this result is that, with the increasing popularity of the Internet in rural areas and the continuous upgrading of rural Internet technology, more and more farmers perceive the usefulness and ease of use of the Internet through cognitive processes, behavior, and self-awareness. However, in addition to the above inherent factors, perceiving risk requires cultivating farmers’ trust. Farmers are willing to adopt new technologies and methods, but the promotion of these technologies takes a while to shorten the psychological distance. Letting farmers know about risk aversion is not enough for farmers to take action.

5. Discussion

This paper constructed a TAM-PR model to explore the effect of Internet use on farmers’ adoption of biopesticides, and 532 rice growers in China’s largest rice-producing area were selected for testing. The results show the following: (1) The use of the Internet can significantly encourage the adoption of biopesticides by farmers, and the application of the Internet affects farmers’ decision-making on the adoption of biopesticides through perceived usefulness and ease of use. The mediating effects were 15.77% and 18.89%, respectively. Adoption decisions with perceived risk without significant impact may be because farmers do not create enough confidence in biopesticides to make bold choices based on Internet knowledge alone. (2) There is heterogeneity in the use of the Internet by farmers to adopt biopesticides. The level of income and the degree of a business’s scale may be limited, and farmers with high incomes and large-scale operations are more willing to use biological pesticides. The effects of low incomes and individual operation were not significant. While educational attainment helps farmers use the Internet and promote biopesticide adoption, the highly educated group shows a stronger adoption intention than the less educated. Third, in the context of the “Internet”, factors such as gender, education level, agricultural technical training, agricultural labor force, urban distance, village infrastructure, pesticide packaging and handling, health awareness, and other factors will affect a farmer’s decision to use biopesticides.

The main contribution of this paper is reflected in three aspects: First, a theoretical framework for the effect of the use of the Internet on the adoption of biopesticide technology by farmers was constructed to successfully link the use of the Internet to the adoption of biopesticides by farmers. In past studies, academia has focused on the influencing factors of biopesticide use by farmers, such as age, income [37], label recognition [38], advisory services [39], and other factors on farmers’ behavior in the adoption of biopesticides. However, little research has been conducted on the differences and constraints of the effect of Internet use on farmers’ adoption of biopesticides in the context of the current digital intelligence era. Cai et al. [40] found that farmers can learn new agricultural technologies and knowledge through the Internet, which motivates them to choose new agricultural production technologies. This is consistent with the findings of this article, in that the use of the Internet can encourage farmers to adopt biopesticides. We hope this study will aid in speeding up network construction and popularize biopesticides in rural China so as to guide more farmers to understand and use biopesticides. Second, we constructed a TAM-PR model to further explain the intrinsic mechanism of Internet use by farmers in order to adopt biopesticides based on perceived usefulness, perceived ease of use, and perceived risk. The Internet can break the information barrier, broaden access to knowledge, and strengthen awareness of green production so that farmers accept biopesticides and use them. In recent years, scholars have conducted a large amount of research on the paradox of farmers’ willingness and behavior [41]. It is believed that farmers’ technology choice behaviors depend mainly on risk preferences [42]. However, our research found that Internet use could influence farmers’ decisions to adopt biopesticide technologies through perceived usefulness and perceived ease of use but not through perceived risk. This may be because farmers do not have enough confidence in biopesticide technology and are not able to boldly choose biopesticides simply by learning through the Internet. This finding complements the existing research. Third, the use of the Internet by farmers to adopt biopesticides has heterogeneity. On the one hand, farmers’ adoption of biopesticides is related to their education level. Farmers with high education levels are more willing to accept new production methods than those with low education levels, which is consistent with the findings of Abdollahzadeh et al. [43]. The reason may be that highly educated farmers are more capable of learning and more aware of environmental protection. The Internet can be better used to acquire new knowledge of agricultural production, and the benefits of biopesticides can be considered for agricultural production in the long term, thus showing a stronger willingness to use it. On the other hand, farmers’ adoption of biopesticides is related to their economic levels and operation scales, which is consistent with the conclusion of Schreinemacher et al. [44]. However, the difference is that Internet use has a significant impact on the adoption of biopesticides for farmers with high incomes and large-scale operations, but not for farmers with low incomes and individual operations. The reason may be that farmers with high incomes are more likely to have access to effective biological control information through the Internet, while farmers with low incomes are less willing to adopt new technologies due to being ill-informed. Farmers with large-scale operations have the advantages of higher income, lower production costs, and strong risk tolerance. They are more adventurous, more receptive to new technologies, more willing to learn about biocontrol technologies through the Internet, and would like to gain excess returns via the adoption of environmentally friendly biopesticides. However, farmers with individual operations have a low ability to resist risk and are more inclined to traditional agricultural production processes and operations, making it difficult for them to accept new technologies even by using the Internet. Even when using the Internet, they struggle to embrace new technologies. This study examines existing research while also proposing new ideas, pointing out a direction for the promotion of biopesticide technology in rural areas of China.

Finally, we would be remiss if we did not acknowledge some of the limitations of the current work, which will be addressed in future research. First, although this article used microdata from the largest grain-producing province in China, macro-statistics at the national level can be used in the future to test the results of this article and further refine the relevant findings. Second, due to the differences in agricultural approaches and desired goals in different countries, the findings of this article are only applicable to developing countries that are making efforts to develop ecological agriculture. However, China has the world’s largest population and is the world’s largest rice producer. The research results are important for China to promote the sustainable development of green agriculture and provide references for some developed countries or less developed regions.

6. Conclusions and Implications

At present, developing countries urgently need to solve the problem of pollution in agricultural production. Especially in a country with a large population, such as China, it is more important to promote environmental sustainability while taking into account agricultural production. Promoting biopesticides instead of chemical pesticides has become the direction for China to promote green agricultural development. In this era of digital intelligence, it is even more necessary to use the Internet to promote the use of biopesticides. This research found that Internet use can effectively promote farmers’ adoption of biopesticides and can influence farmers’ decisions through perceived usefulness and perceived ease of use. There is heterogeneity in Internet use by farmers with different education levels, income levels, and operation scales with respect to their adoption of biopesticides. This study enriches the research on the green production behavior of farmers. Considering China’s gradual development toward digital agriculture, it is particularly important to guide farmers to adopt biopesticides through the Internet. The conclusion can help the government carry out rural construction, guide farmers to adopt pro-environmental production technologies, and promote the sustainable development of agriculture in China.

Based on the above analysis, the following policy implications are proposed in this article: First, we should promote the construction of rural Internet technology infrastructure; strengthen the construction of rural hard Internet infrastructure; promote the construction of rural gigabit networks; and increase and reduce the cost of rural network broadband by introducing 5G technology, remote sensing satellites, mobile Internet, and other space-based infrastructure. We should also promote the construction of digital demonstration villages, guide and train farmers to use the Internet, and inspire farmers to use the Internet to acquire advanced agricultural production knowledge. Second, we should innovate the Internet biopesticide promotion system. Taking advantage of the regional network characteristics of the Internet, we can integrate scattered organizations such as cooperatives, agricultural extension stations, and agricultural dealers in provinces, strengthen unified training and upgrades, and encourage agricultural extension stations to create portals and APP promotion platforms to facilitate the instant and regular transmission of biological pesticides. Varieties and application methods can be introduced to farmers through the Internet, which lays a grassroots foundation for the promotion of biological pesticides. Third, we should give full play to the important role of the Internet in promoting the adoption of biopesticide technology to upgrade farmers. More attention should be paid to the differentiation trend in farmers’ technology adoption behaviors brought about by Internet use. Farmers with different resource endowments should choose appropriate biopesticide ratios and varieties to bridge the gap between groups and broaden the scope of biopesticides.