Front. Sustain. Food Syst. Frontiers in Sustainable Food Systems Front. Sustain. Food Syst. 2571-581X Frontiers Media S.A. 10.3389/fsufs.2025.1611347 Sustainable Food Systems Original Research Impact of meteorological disaster shocks and collective action on farmers' adoption of soil and water conservation technologies: evidence from the Loess Plateau, China Ding Xiuling 1 * Li Lipeng 2 Lu Qian 3 Zhang Weiping 1 1College of Economics, Gansu University of Political science and Law, Lanzhou, China 2School of Economics and Management, Ningxia University, Yinchuan, China 3College of Economics and Management, Northwest A&F University, Xianyang, China

Edited by: Siyabusa Mkuhlani, International Institute of Tropical Agriculture (IITA), Kenya

Reviewed by: Inoussa ZAGRE, Université des Sciences, des Techniques et des Technologies de Bamako, Mali

Alfred Tunyire Apio, University of Pretoria, South Africa

 *Correspondence: Xiuling Ding dxl7482@gsupl.edu.cn 18 09 2025 2025 9 1611347 14 04 2025 25 08 2025 Copyright © 2025 Ding, Li, Lu and Zhang. 2025 Ding, Li, Lu and Zhang

This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

 Introduction

Highlighting the importance of addressing the meteorological disaster shocks (MDS) experiences of potential adopters, alleviating disaster-induced anxiety, and harnessing risk aversion psychology, along with promoting the essence of collective action, can be crucial in facilitating the adoption of soil and water conservation technologies (SWCT).

 Methods

This study delves into the intricate dynamics governing the adoption of SWCT among grain growers. This study draws on field survey data from 1,106 farming households across three provinces. Using a binary Probit model and mediation effect analysis, it explores how MDS and collective action influence farmers' adoption of SWCT.

 Results

The results reveal several key findings. First, MDS significantly and positively affect farmers' adoption of different types of SWCT. Second, collective action also has a significant positive impact on SWCT adoption. Third, collective action mediates the relationship between MDS and SWCT adoption, indicating that part of MDS's influence is transmitted through collective action. Finally, heterogeneity analysis shows that the impact of MDS on SWCT adoption varies significantly across farmer subgroups defined by age, education level, and planting scale. Likewise, government subsidies, total household land management area, and the number of household agricultural machinery significantly affect farmers' decisions to adopt SWCT.

 Discussion and suggestions

Based on the research, the following suggestions are put forward: promote adaptive technologies and strengthen demonstrations and guidance; improve meteorological early warning, establish a disaster case database, and subsidize disaster-stricken farmers; support cooperatives in participating in technology promotion and reward excellent organizations; for farmers aged 55 and below with low educational levels, promote simple technologies, strengthen training, increase subsidies, and drive adoption through demonstrations; optimize subsidy policies, allocate technologies based on resources, and establish an evaluation mechanism for dynamic policy adjustment.

 meteorological disaster shocks collective action soil and water conservation technology mediation effect heterogeneity analysis 71873102 National Natural Science Foundation of Chinahttps://doi.org/10.13039/501100001809 section-at-acceptance Agroecology and Ecosystem Services Introduction

The impacts of climate change are increasingly intensifying, characterized by rising temperatures, erratic rainfall patterns, and increasingly frequent extreme weather events, which have significantly heightened the vulnerability of agricultural systems (Parry, 2019). These vulnerabilities, in turn, have introduced substantial uncertainty into farmers‘ incomes and exerted direct and indirect influences on long-term national food security and broader social and economic development agendas (Thornton et al., 2014). In mitigating such negative impacts, soil and water conservation technologies (SWTC) have emerged as a crucial solution. SWTC mainly fall into three categories: engineering technologies (such as terrace drainage, contour planting, and fish-scale planting) (Fontes, 2020; Li et al., 2021), biological technologies (such as shelterbelt construction and grass planting), and tillage technologies (such as no-tillage, reduced tillage, and straw mulching) (Jia and Lu, 2018; Huang et al., 2019). They play roles in conserving soil and water, improving rainfall infiltration and soil structure, increasing soil organic matter content, reducing soil erosion, enhancing land productivity, and improving the ecological environment (Kpadonou et al., 2017), thus ensuring farmers' incomes. Meanwhile, they are regarded as important components in achieving the United Nations Sustainable Development Goals, capable of delivering triple benefits—economic, ecological, and social (Altobelli et al., 2020; Ojo et al., 2021)—and serving as effective technologies for resisting disaster risks (Ostrom et al., 1979).

By 2020, over 35% of countries worldwide had promoted and adopted SWTC, covering 11% of the world's arable land. For example, more than 60% of arable land in countries like the United States and Canada had adopted such technologies (Cui et al., 2023). Since the 1940s, the Chinese government has implemented a series of incentive policies and measures to encourage farmers to adopt SWTC, and the application level of various technologies has made certain progress. By 2020, China had built 376,800 hectares of basic farmland using engineering measures, created 1,410,900 hectares of soil and water conservation forests and planted 398,300 hectares of grass using biological measures (Hu et al., 2023; Lei et al., 2023), with soil conservation tillage measures implemented on 1,461,700 hectares of land. However, it should be emphasized that there are many problems in the promotion and application of SWTC. the adoption level of some SWTC is far lower than the international level, especially in the ecologically fragile Loess Plateau region of China (Cao et al., 2017; Jiang et al., 2019; Zhang et al., 2023). farmers, as implementers and beneficiaries of SWCT adoption, have low enthusiasm for active adoption and a relatively low continuous adoption rate, making it difficult for these technologies to effectively mitigate the negative impacts of climate change and MDS (Lingling et al., 2014; Jia and Lu, 2019; Gu et al., 2021; Chen et al., 2023). Therefore, in-depth research and analysis of the determinants influencing farmers‘ adoption behavior of SWTC are of great significance for formulating effective measures to promote farmers' adoption of such technologies.

In exploring how to encourage farmers to actively adopt SWTC, the academic community has conducted in-depth discussions mainly from two aspects: internal factors and external environment (Teshome et al., 2016; Jia and Lu, 2019). Regarding internal factors, relevant studies have found that household head characteristics (such as gender, age, and education level) and household endowment characteristics (such as social capital, human capital, agricultural income, and land management area) significantly affect farmers‘ adoption of SWTC (Wei C. et al., 2022; Wang et al., 2023). in terms of the external environment, some scholars have focused on analyzing the external environmental factors of implementers and found that land property rights, government support (technology promotion, technical training, and subsidies), and regional economic development level all influence farmers' adoption of SWTC (Danso-Abbeam, 2022; Wang Y. et al., 2022). It is worth noting that existing studies have analyzed farmers‘ adoption of Soil and Water Conservation Technologies (SWCT) from both internal and external influencing factors of farm households, which is of important reference value. However, these studies lack attention to farmers' experiences with meteorological disasters.

The study adopts imprinting theory to reveal the impact of meteorological disaster shocks on farmers‘ cognition and behavior (Avital and Jablonka, 1994). From the farmers' perspective, meteorological disaster shocks (MDS) refer to the direct or indirect impacts of meteorological disasters on their production, income, and lives. Since SWTC have the function of resisting natural disaster risks, MDS are closely related to farmers‘ adoption behavior of SWTC. Farmers' decisions on the use and management of natural resources largely depend on their perception and sensitivity to past experiences. After being impacted by meteorological disasters, farmers adjust their behaviors and measures according to their own circumstances, continuously learn, and proactively adopt SWTC to better cope with future threats (Teshome et al., 2016). MDS will be imprinted in farmers‘ and collective consciousness, affecting the decision-making and practices of agricultural communities. Moreover, the implementation process and effects of SWTC can improve the surrounding ecological environment, benefit the agricultural production environment of other farmers, and have the attributes of positive externality and non-exclusivity, being typical public goods. Therefore, when studying the relationship between MDS and farmers' adoption behavior of SWTC, it is necessary to consider the role of collective action.

To sum up, there are many research results on farmers' adoption behavior of SWTC, which are of important reference value. However, few studies have analyzed farmers' adoption behavior of such technologies from the perspective of meteorological disaster, and even fewer have integrated MDS, collective action, and farmers' adoption behavior into a unified analytical framework, which provides room for further research in this paper. In view of this, this paper constructs an analytical framework involving MDS, collective action, and farmers' adoption behavior of SWTC. Using survey data of farmers from Shaanxi, Gansu, and Ningxia provinces, it examines and analyzes the impacts of MDS and collective action on farmers' adoption behavior of SWTC, and further explores the differential impacts of MDS on the adoption of such technologies by different types of farmers. The research significance lies in first, it helps improve the theoretical research system of farmers' behavior, providing a new perspective for studying the determinants of adoption behavior of SWTC from the perspectives of MDS and collective action. Second, based on the imprinting theory, exploring the impact of MDS on farmers' choice of SWTC can enrich the imprinting theory. Third, starting from the characteristics of technology adopters, it can more effectively provide a scientific decision-making basis for policies related to promoting SWTC.

The rest of the paper is structured as follows: Section 2 elaborates on the theoretical analysis framework and research hypotheses, clarifying the logical relationships between key variables; Section 3 explains the research area, data sources, variable descriptions and statistics, and the setting of empirical models; Section 4 presents and analyzes the empirical results, including benchmark regression, mediating effect test, and heterogeneity analysis; Section 5 contains the research conclusions, policy recommendations, and discussions.

 Theoretical analysis and hypotheses

The theory of imprinting, originally grounded in the field of biology and subsequently incorporated into the realms of psychology and behavioral science, delves into the concept of enduring individual experiences that remain resilient over time, unaltered by shifting environmental conditions (Marquis and Tilcsik, 2013). This theory wields a profound and persistent influence on an individual's subsequent decision-making processes, affecting their propensity for risk aversion, cognitive learning awareness, and preventive cognition. In the context of farmers' encounters with disasters and their collective responses, imprinting theory offers valuable insights into the intricate mechanisms underpinning resilience and adaptability within agricultural communities (Wigboldus et al., 2016).

A crucial aspect to consider is the expeditious and involuntary acquisition of knowledge. Farmers who have weathered catastrophes such as devastating floods, prolonged droughts, or the emergence of invasive pests undergo a parallel process of acquiring essential skills (Wang et al., 2015; Xie et al., 2021). They acquire the capacity to swiftly recognize early warning signs, adapt their agricultural methodologies, and institute effective measures to mitigate losses and secure their livelihoods (Lin and Chang, 2020). These farmers, equipped with the knowledge and skills necessary for proficient disaster management, frequently convene to exchange their experiences and insights (Thapa et al., 2022). Likewise, collective action nurtures a sense of community and mutual support. It ultimately reinforces the overall resilience of the agricultural community (Ombogoh et al., 2018). It plays an indispensable role in shaping cognitive frameworks and establishing communal norms (Qurniati et al., 2017). The process of imprinting disaster management experiences results in the accumulation of shared knowledge and the development of a collective ethos. This, in turn, significantly influences decision-making processes and the formulation of preparedness plans within the farming community (Belachew et al., 2020). Farmers collaborate to modify their agricultural techniques, conceive more robust strategies for disaster management, and allocate resources for the construction of infrastructure benefiting the entire community. Given the mounting environmental challenges faced by agricultural communities, acknowledging the critical significance of imprinting theory is imperative in devising effective strategies to enhance disaster resilience (Takayama et al., 2018; Villamayor-Tomas et al., 2019).

The impact of MDS on the adoption behavior of farmers' SWCT

The study adopts imprinting theory to depict the profound impacts of meteorological disaster shocks on farmers' cognition and behavior. It emphasizes how such shocks influence farmers' cognitive structures and decision-making behaviors in the process of agricultural production. Based on existing research, the analysis of the mechanism of farmers‘ MDS on their adoption of SWCT includes the following three aspects (Rozaki et al., 2021; Fusco et al., 2021): First, farmers who have experienced MDS will increase their risk aversion and enhance their awareness and ability to resist risks (Ullah et al., 2015). Risk aversion to current or future disasters prompts farmers to adopt risk-avoiding strategies. At the same time, SWCT is a MDS-resistant technology that can reduce future ecological and environmental risks, increase agricultural income, and reduce farmers' environmental vulnerability and livelihood vulnerability (Bewket, 2007). Therefore, farmers who have experienced MDS are more inclined to adopt soil and water conservation technologies and then, reduce the current or future risk exposure level of farmers (Belayneh, 2023).

Second, farmers who have experienced MDS will actively enhance their ability to resist MDS by adopting SWCT. Specifically, after being impacted by meteorological disasters, farmers will change their learning awareness, actively acquire and learn information and technologies related to soil and water conservation, thereby strengthening household human capital and material capital. Ultimately, farmers will implement SWCT to prevent and mitigate the negative impacts of future MDS on household agricultural production or household livelihoods.

Third, in China, most people engaged in agricultural and forestry production are middle-aged and elderly, with a strong dependence on land. Moreover, this group of farmers has experienced more MDS (Jarray et al., 2023; Ogunniyi et al., 2023), which deepens their memories of the losses and risks caused by MDS and induces psychological panic about disaster losses (Qin et al., 2020; Indrawati et al., 2022). Therefore, to reduce farmers' risk exposure levels and potential losses caused by future MDS (Huo et al., 2020), farmers tend to adopt risk avoidance strategies, namely SWCT.

 The impact of collective action on the adoption behavior of farmers' SWCT

In 1979, Ostrom et al. (1979) revised and developed Olson's collective action theory and proposed a method to solve the collective action dilemma and explained that an efficient system promotes mutual supervision, and strengthening mutual supervision can ensure that people make credible commitments (Ostrom and Ostrom, 2019). The three mutually restrict and complement each other. At the same time, Jara-Rojas et al. (2012) believe that collective action can encourage farmers to adopt SWCT. Collective action in this paper refers to farmers' active and passive participation; a sound regulatory mechanism established to achieve common interests or goals—distributing meteorological disaster shock risks and sharing disaster management costs (Liu et al., 2013); an appropriate information transmission mechanism for exchanging information and technologies; collective organizational activities for negotiation (Yang et al., 2005; Huang et al., 2020a).

On the one hand, formal and informal supervision mechanisms and information transmission mechanisms within collective organizations have positive effects on farmers' adoption of soil and water conservation technologies (Wang et al., 2022). Participating in collective action can promote mutually beneficial cooperation between farmers and can effectively learn and acquire new technology and knowledge information in formal collective activities or mutual exchanges among collective organization members, resulting in a “learning by doing” effect (Bizoza, 2014; Pang et al., 2020). Under the supervision of the action organization system and the mutual supervision of members of collective organizations, it is possible to standardize the technical use standards of farmers, improve the quality of agricultural production, and encourage farmers to pay attention to the family income while taking into account the awareness and behavior of ecological environment protection, thereby increasing the adoption rate of soil and water conservation technologies (Dennis and Brondizio, 2020; Ombogoh et al., 2018).

On the other hand, collective action promotes farmers' adoption of soil and water conservation technologies by sharing the risks and costs incurred in the adoption process (Nigussie et al., 2018). The members of the collective organization collectively produce labor, and the unified purchase or cooperative purchase of soil and water conservation technical equipment and agricultural materials can reduce or share the cost of technology adopters (VanOel et al., 2019). Through collective purchase, the bargaining power will be improved, which will reduce the cost of information search and information acquisition and reduce the cost of individual farmers. Transaction costs and risks, realize the interests of members participating in collective action, generate demand-induced effects, form a good mechanism of interest linkage, and promote farmers to adopt SWCT (Gebremeskel et al., 2018; Assefa et al., 2021). However, collective action can actively promote the adoption of soil and water conservation technologies by farmers under the influence of the collective organization's supervision mechanism, information transmission mechanism, and cost-sharing mechanism (Gebremeskel et al., 2018).

 The influence mechanism of MDS and collective action on the adoption behavior of farmers' SWCT

This paper adopts a public goods game to discuss how MDS changes individual learning awareness, cooperation awareness, and risk awareness, which in turn leads to prosocial behavior (participation in collective action), diversification of risks or risk losses, and sharing of risk-taking costs (Nyangena, 2008), thereby promoting farmers to adopt SWCT as public goods (Streletskaya et al., 2020). In reality, after farmers have experienced the impact of MDS, they will become averse to current or future disaster risks and induce panic about past risk losses (Brown et al., 2018; Budhathoki et al., 2020). They will actively establish a restoration and protection mechanism to restore agricultural production. An important approach and means to avoid panic over post-disaster losses and restore agricultural production after disasters is to obtain external funds, technologies, and information. This gives rise to the phenomenon of “huddling together for warmth”. It further promotes the implementation of SWCT, and reduces farmers' risk exposure levels and the probability of MDS (Darkwah et al., 2019; Sileshi et al., 2019).

First, when farmers experience the impact of MDS, participating in village collective activities, cooperatives, and cooperative supply organizations in the village will obtain shared benefits brought about by the rules and regulations set up by the organization (Mekuriaw et al., 2018; Wordofa et al., 2020). During activities, they will take the initiative to strengthen contact with other farmers in the same village by participating in collective actions and overcoming the panic caused by the impact of disasters (Nkegbe and Shankar, 2014; Fallah et al., 2016). Farmers will seek help, whether it is agricultural production materials or emotions, and then strengthen information and technology (Fallah et al., 2016). Communication to increase the adoption rate of farmers' soil and water conservation technologies. Second, when farmers who have experienced the impact of MDS participate in cooperatives, companies, village collective activities, and cooperative supply organizations in the village in order to restore agricultural production, the organization will provide professional technical guidance (Belayneh et al., 2019; Morris and Arbuckle, 2021) and accurate information based on the characteristics of soil and water conservation technologies (Batjes, 2014). Moreover, as a part of the capital investment, they also share the cost of farmers' technology adoption, and increase farmers' behavior of SWCT adoption (Chesterman et al., 2019). Based on the above theoretical analysis, a theoretical analysis framework diagram (Figure 1) has been constructed.

Theoretical analysis framework diagram of this study.

Flowchart illustrating the relationship between meteorological disaster shocks, collective action, and soil and water conservation technology adoption. Arrows show the pathway from disaster shocks to collective action, leading to technology adoption. Three hypotheses include: Hypothesis 1, reducing risk exposure levels to decrease disaster occurrence; Hypothesis 2, diversifying risk management costs; and Hypothesis 3, mediating through collective action. Materials and methods Data source and sample description

The research data in this paper comes from the questionnaire survey conducted by the research group on grain growers in Northwest China from October to November 2016. Shaanxi Province, Gansu Province, and Ningxia Hui Autonomous Region are all located in the northwest region, belonging to the Loess Plateau area and the irrigation area in the upper and middle reaches of the Yellow River Basin (see Figure 2 Geographical location of the study area and distribution of counties in the field survey.). They are representative areas for key control of soil erosion under the corn and wheat crop planting system. Yang District, Mizhi County and Suide County, Yuanzhou District and Pengyang County in Guyuan City, Ningxia Hui Autonomous Region, and Xifeng District and Huan County in Qingyang City, Gansu Province were used as research sites. In most areas of these three provinces, the soil is loose, the terrain is fragmented, vegetation coverage rate is low and the climate is arid. Not only is soil erosion a severe problem here, but meteorological disasters such as droughts, floods, hailstorms, and sandstorms also occur frequently. These conditions lead to a fragile ecological environment and a decline in crop yields. Therefore, farmers in this area, with both representativeness and diversity, are appropriate survey respondents. This survey adopts a combination of typical sampling and random sampling methods. Specifically, 2–3 counties were selected from each of Shaanxi Province, Gansu Province, and Ningxia Hui Autonomous Region; 3–5 towns were chosen from each county; 2–5 villages were selected from each town; and 20 respondents with communication skills were randomly selected from each village. Finally, after processing outliers and missing values, a total of 1,106 valid samples were obtained, including 334 from Shaanxi Province, 379 from Gansu Province, and 393 from Ningxia Hui Autonomous Region.

Geographical location of the study area and distribution of counties in the field survey.

A map showing data collection areas in China, highlighting Gansu, Shaanxi, and Ningxia provinces in blue. Each province's map is expanded to indicate specific locations with red triangles: Huanzian and Xifeng in Gansu; Yuyang, Mizhi, and Suide in Shaanxi; Yuanzhou and Pengyang in Ningxia. A legend indicates map colors and symbols.

Table 1 shows the adoption of SWCT by sample farmers in the survey area. It can be found that among the sample farmers, 516 households adopted engineering technology, 141 households adopted biotechnology, and 247 households adopted farming technology, accounting for 46.67%, 12.75%, and 22.33% of the total number of sample farmers, respectively; at the same time, Ningxia farmers adopted the most SWCT, followed by Gansu and Shaanxi.

Adoption of SWCT by sample farmers in the survey area.

Province Engineering technology Biotechnology Farming technology
Number of households Proportion (%) Number of households Proportion (%) Number of households Proportion (%)
Gansu 135 12.21 24 2.17 92 8.32
Shaanxi 91 8.23 81 7.32 21 1.90
Ningxia 290 26.22 36 3.25 134 12.12
total 516 46.65 141 12.75 247 22.33

According to the basic characteristics of the sample farmers (as shown in Table 2), 61% of the respondents are under 55 years old, 56% have received education at or above the junior high school level. Judging from the family land management scale of sample farmers, the proportion of farmers with more than 30 mu reaches 13.20%. From the perspective of the proportion of non-agricultural income, the non-agricultural income of the sample farmers accounted for more than 75% and accounted for 54.07% of the total sample size. Compared with the eastern and central regions with better ecological environments and developed economies, the farmers in the sample areas show a small scale of family land management, and the main source of household income is non-agricultural income.

Distribution characteristics of household heads' age and educational level among sample farmers.

Variable name Classification Frequency Percentage (%) Variable name Classification Frequency Percentage (%)
Age Under 45 years old 261 23.60 Land management scale 3 acres and below 99 8.95
46–55 years old 411 37.16 3–10 acres 392 35.44
56–65 years old 283 25.59 10–30 acres 469 42.41
Over 66 years old 151 13.65 30 acres 146 13.20
education level Illiteracy 186 16.82 Proportion of non-farm income 25% and below 196 17.72
Primary school 438 39.60 25–50% 110 9.94
Junior high school 359 32.46 50–75% 202 18.27
High school and above 123 11.12 More than 75% 598 54.07
Variable selection and statistical description Explained variables

This article selects the core technology of SWCT, including engineering technology, biotechnology, and farming technology. When analyzing the adoption behavior of farmers' SWCT, this paper refers to the method of Mao et al. (2024). The “variable” is set as a dummy variable, 0 means that farmers do not adopt a certain SWCT, and 1 means that farmers have a certain SWCT adoption behavior (as shown in Table 3).

Variable definitions, assignments, and statistical descriptions of sample farmers.

Category Variable name Variable definitions AV SD
Explained variable Farmers' SWCT adoption behavior Engineering technology adoption behavior Whether farmers have adopted engineering technologies such as building terraces, terraces, and fish scale pits (yes = 1, no = 0) 0.4665 0.4991
Biotechnology adoption behavior Whether farmers have adopted biotechnology such as planting trees and grasses (yes = 1, no = 0) 0.1274 0.3337
Farming technology adoption behavior Whether farmers have adopted farming techniques such as low-tillage and no-tillage, furrow and ridge farming (yes = 1, no = 0) 0.2 233 0.4166
Core explanatory variable MDS whether farmers experienced many droughts, rainstorms, floods, strong winds and hailstones from 2014 to 2016 (yes = 1, no = 0) 0.3376
Mediator variable Collective action whether to participate in cooperatives, companies, collective activities in the village, spontaneous collective organizations in the village (yes = 1, no = 0) 0.4612
Control variable Personal characteristics of the head of household Age Actual age of head of household (years) 53.1066 10.9697
Education level Education level of head of household (years) 5.8318 3.7196
Technology ecological value cognition The improvement degree of SWCT to the ecological environment of the area (no effect = 1, some effect = 2, general = 3, large effect = 4, very large effect = 4) 0.3943 0.1824
Family characteristics The proportion of male labor force Number of the male labor force in household/total labor force (%) 16.6745 15.7953
Land management area Total land area owned by households (mu) 0.0958 0.2945
Village cadres Whether the family member has a village cadre (yes = 1, no = 0) 0.4529 0.5692
Number of farm machinery The number of agricultural machinery owned by rural households (pieces) 0.6694 0.3364
Proportion of non-farm income Household off-farm income/total income (%) 0.6238 0.4846
Social network Number of frequent family contacts (number) 61.4719 89.1034
External environment characteristics Government subsidies Whether the government has subsidies for SWCT (yes = 1, no = 0) 1.6031 0.8846
Gansu Is it located in gansu (yes = 1, no = 0) 0.3426 0.4748
Ningxia Is it located in ningxia (yes = 1, no = 0) 0.3553 0.4788
Core explanatory variables and mediator variables

The core explanatory variable selected in this paper is MDS. The core explanatory variable selected in this paper is MDS. Referring to the study by Peng et al. (2020), MDS are treated as a dummy variable. The variable is assigned a value of 1 if the respondent has experienced or suffered MDS. The variable is assigned a value of 0 if the respondent has not. Additionally, collective action is selected as another core explanatory variable and mediating variable. Collective action mainly refers to whether farmers choose to participate in collective organizations such as cooperatives, companies, village collective activities, and spontaneous and autonomous cooperative supply organizations within the village during agricultural production, so as to spread risks, share agricultural production costs, and other related matters. This variable of collective action is also treated as a dummy variable. If the interviewed farmer participates in collective action, the variable is assigned a value of 1. If the interviewed farmer does not participate in collective action, the variable is assigned a value of 0 (as shown in Table 3).

 Control variables

According to the existing related research (Huang et al., 2020b; Upadhaya et al., 2023), this paper selects other factors that affect the adoption -behavior of farmers‘ SWCT as control variables, including the personal characteristics of the head of household (age, education level, cognition of technological, ecological value), family characteristics (proportion of male household labor force, area of land management, number of household agricultural machinery, the proportion of non-agricultural income, whether there are village cadres among family members, and social network) and external environmental characteristics (government subsidies and geographical location). Additionally, the variable of the proportion of male labor force is characterized by the ratio of the number of male laborers to the total number of laborers in the family labor force. The variable of the proportion of non-agricultural income is characterized by the ratio of non-agricultural income to the total household income in family income. The variable of social network is referenced from the research of Ashoori et al. (2016), and is characterized by the number of people with whom the family frequently interacts. The variable of Technology ecological value cognition refers to farmers' cognition of the role of (SWCT) in improving the ecological environment, is measured using a five-level Likert scale.

 Model setting Baseline model

The explained variable is farmers' adoption behavior of soil and water conservation technologies, which specifically includes whether farmers have adopted engineering technologies, biological technologies, or tillage technologies. These are all 0–1 variables, so it is reasonable to use the binary Probit model for regression analysis. The benchmark model set in this paper is as follows:

Yi = a0+a1Xi+a2C+δ

In equation (1), Y is the explained variable, indicating the adoption behavior of farmers' SWCT, X indicates whether the farmers have experienced MDS or participated in collective actions, C indicates other control variables, and αi indicates the coefficient to be estimated, which δ is a random error item.

 Mediation effect model

To test the intermediary effect, we use the intermediary effect model proposed by Wen and Ye (2014) to analyze the mechanism of action among MDS, collective action, and farmers' SWCT adoption behavior. The model settings are as follows:

Y = a0+a1X+a2C+δ M = β0+β1X+β2C+ε Y = γ0+γ1M+γ2X+γ3C+ϑ

In equations (2)–(4), Y is the explained variable, representing farmers‘ adoption behavior of SWCT; X denotes whether farmers have experienced MDS; the mediating variable M indicates whether farmers participate in collective action; C indicates the control variable, while αi, βi, γi are the coefficients to be estimated. Specifically, Equation (2) is used to test the impact of MDS on farmers' adoption behavior of SWCT. Equation (3) examines the influence of MDS on farmers‘ participation in collective action related to SWCT adoption; Equation (4) incorporates both MDS and collective action to verify their combined effect on farmers' adoption behavior of SWCT. The steps for testing the mediating effect in this paper are as follows: First, test whether the coefficient α1 of Equation (2) is significant. If it is significant, proceed to the next step. Second, sequentially test the coefficient β1 of Equation (3) and the coefficient γ1 of Equation (4). If both are significant, the mediating effect is significant, and proceed to the fourth step. If at least one is not significant, proceed to the third step. Third, directly test H0: β1γ1 = 0 using the Bootstrap method. If significant, the mediating effect is significant, and proceed to the fourth step; otherwise, the mediating effect is not significant, and stop the analysis. Fourth, test the coefficient γ2 of Equation (4). If it is not significant, it indicates that there is only a mediating effect. If it is significant (i.e., the direct effect is significant), proceed to the fifth step. Fifth, compare the signs of β1γ1 and γ2. If they have the same sign, it is a partial mediating effect.

 Results

Before the formal regression analysis of the benchmark model, the correlation coefficient between the variables was preliminarily tested, and the correlation between MDS, collective action and the adoption behavior of farmers' soil and water conservation technologies was tested more intuitively (see Table 4). It can be seen from Table 4 that there is an obvious positive correlation between MDS and collective action and the adoption of SWCT by farmers, which is consistent with our theoretical analysis. At the same time, considering the possible multi-collinearity among the variables, VIF (variance inflation factor) was used for testing. The test results showed that the maximum VIF value was 1.36, and the VIF values of all variables were less than 10. The variables are independent of each other, and there is no multi-collinearity problem that affects the regression results.

Correlation analysis results between core explanatory variables and explained variables.

Variable name Engineering technology Biotechnology Farming technology MDS Collective action
Engineering technology 1.0000 — — — —
Biotechnology 0.1316*** 1.0000 — — —
Farming technology 0.3036*** 0.1595*** 1.0000_ — —
MDS 0.1270*** 0.1164*** 0.1054*** 1.0000 —
Collective action 0.1736*** 0.1659*** 0.1069*** 0.2240*** 1.0000

*, **, and *** indicate significance at the statistical level of 10%, 5%, and 1 %, respectively.

 Analysis of benchmark model regression results

The estimated results are shown in Table 5. It can be seen from the regression results that the Wald test values are all significant at the 1% statistical level, indicating that the setting of the model is statistically effective and the overall fit of the model is good.

Empirical regression results of the impact of MDS and collective action on farmers' adoption behavior of SWCT.

Variable name Engineering technology Biotechnology Farming technology Engineering technology Biotechnology Farming technology
Model (1) Model (2) Model (3) Model (4) Model (5) Model (6)
MDS 0.3176** (0.1285) 1.0953*** (0.2434) 0.3109* (0.1604)
Collective action 0.5252*** (0.0930) 0.6731*** (0.1447) 0.3218*** (0.1047)
Age 0.0029 (0.0041) 0.0010 (0.0054) 0.0055 (0.0046) 0.0031 (0.0041) 0.0009 (0.0053) 0.0052 (0.0046)
Education level −0.0140 (0.0114) 0.0376** (0.0147) 0.0398*** (0.0126) −0.0159 (0.0115) 0.0351** (0.0144) 0.0388*** (0.0126)
The proportion of male labor force 0.2648 (0.2278) 0.2524 (0.2924) 0.0285 (0.2524) 0.2248 (0.2289) 0.1916 (0.2943) 0.0125 (0.2546)
Land management area 0.0213*** (0.0034) 0.0140*** (0.0034) 0.0080*** (0.0029) 0.0211*** (0.0035) 0.0128*** (0.0034) 0.0077*** (0.0029)
Village cadres −0.0583 (0.1476) 0.0644 (0.1794) 0.0352 (0.1543) −0.0978 (0.1494) 0.0172 (0.1785) 0.0241 (0.1544)
Number of farm machinery 0.0929 (0.0761) 0.1109 (0.0967) 0.3193*** (0.0797) 0.0995 (0.0769) 0.1249 (0.0979) 0.3181*** (0.0799)
Proportion of non-farm income 0.0052 (0.1300) 0.2491 (0.1719) −0.0473 (0.1431) −0.0495 (0.1312) 0.0904 (0.1718) −0.0775 (0.1440)
Government subsidies 0.2850*** (0.0882) 0.1448 (0.1211) 0.0088 (0.0969) 0.2437*** (0.0893) 0.0855 (0.1223) −0.0169 (0.0975)
Social network 0.0002 (0.0005) −0.0004 (0.0006) −0.0007 (0.0006) 0.0001 (0.0005) −0.0006 (0.0006) −0.0009 (0.0006)
Technology ecological value cognition 0.0392 (0.0494) 0.2236*** (0.0603) 0.2178*** (0.0525) 0.0236 (0.0498) 0.2452*** (0.0599) 0.2052*** (0.0528)
Gansu (with shaanxi as the reference group) 0.1163 (0.1108) −1.2119*** (0.1520) 0.6836*** (0.1414) 0.1889* (0.1114) −1.1347*** (0.1532) 0.7347*** (0.1425)
Ningxia (with shaanxi as the reference group) 1.0204*** (0.1154) −0.9206*** (0.1502) 1.0575*** (0.1461) 1.0926*** (0.1160) −0.8184*** (0.1486) 1.0909*** (0.1461)
Constant term −1.5791*** (0.3299) −2.8193*** (0.4694) −2.8783*** (0.3883) −0.5409* (0.2789) −3.1454*** (0.3850) −1.4782*** (0.3047)
Prob > chi2 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
Log likelihood −627.3150 −352.6724 −505.6786 −614.1264 −354.3611 −502.8051
Pseudo r2 0.1791 0.1643 0.1391 0.1963 0.1603 0.1440
Wald test value 232.67*** 113.73*** 132.36*** 248.25*** 108.45*** 138.52***

*, ** and *** indicate significance at the statistical level of 10%, 5%, and 1%, respectively.

Models (1), (2), and (3) in Table 5 are the regression results of MDS on farmers' adoption of SWCT. In Model (1), the coefficient for MDS is positive and statistically significant at the 5% level. This indicates that such experience has a positive impact on farmers' adoption behavior of engineering technologies. In Model (2), the coefficient for MDS is positive and statistically significant at the 1% level. This suggests that MDS has a positive impact on farmers' adoption behavior of biological technologies. In Model (3), the coefficient for MDS is positive and statistically significant at the 1% level. This indicates that MDS has a positive impact on farmers' adoption behavior of farming technologies. Based on the above analysis, it is explained that: farmers who have experienced MDS are more inclined to adopt SWCT, the reason is to avoid the risk of MDS, reduce their own risk exposure level, improve the ecological environment of the area and reduce their own risk. The vulnerability of livelihoods promotes farmers to adopt MDS risk-resisting technologies to increase the adoption rate of soil and water conservation technologies. Hypothesis 1 has been verified. Table 5 outlines the regression results of models (4), (5), and (6), covering the collective action on farmers' adoption of SWCT. The impacts are all significant at the 1% significance level, and the coefficients are positive; it shows that farmers' participation in collective action has a positive effect on farmers' adoption of engineering technology, biotechnology and farming technology in soil and water conservation technologies, because in reality, soil and water keeping technology with the attribute of public goods, in order to avoid the phenomenon of “free riding”, farmers often realize the supply of public goods through institutional supervision mechanisms of collective action, information communication mechanisms, and cost-sharing mechanisms, that is, to promote the adoption of soil and water conservation technologies by farmers. Behavior, hypothesis 2 was verified.

Among the control variables, the coefficient for the household head's education level is positive, with significance at the 5% and 1% levels respectively. This indicates that farmers with higher education levels are more inclined to adopt these two types of technologies. The potential reason is that such farmers, with higher education, can not only better understand and apply biological and farming technologies but also have a clearer awareness of the risk-resistant functions of these technologies, thus being more willing to adopt them. The coefficient for land management area is positive and significant at the 1% level across all cases. This indicates that the larger the land management area, the more willing farmers are to adopt SWCT. There are two possible reasons: On one hand, to expand their land management area, farmers use engineering, biological, and farming technologies to convert wasteland and severely desertified land into usable land, thereby promoting the adoption of SWCT. On the other hand, farmers with larger land management areas are more concerned about huge losses caused by MDS risks, so they actively adopt SWCT to avoid such losses.

To avoid MDS, large-scale farmers are more inclined to adopt biotechnology and farming technology. Additionally, the coefficient for the number of agricultural machineries is positive and statistically significant at the 1% level. This indicates that the more agricultural machineries there are, the more inclined farmers are to adopt farming technology. A possible reason is that a greater amount of agricultural machinery makes agricultural production methods more convenient and labor-saving, which in turn makes farmers more inclined to adopt farming techniques. Furthermore, the coefficient for technology ecological value cognition is positive and statistically significant at the 1% level. This indicates that the clearer farmers‘ cognition of the ecological value of SWCT is, the more it can promote their use of biotechnology and farming techniques. Moreover, the impact of government subsidies on engineering technology adoption behavior is significant at the 1% significance level, with a positive coefficient, indicating that government subsidies can actively promote farmers' adoption of engineering technology, which is consistent with the research results of Huang et al. (2020c). In addition, the regional dummy variables, taking Shaanxi Province as the reference group, have more significant estimation results, indicating that there are significant regional differences in the adoption of soil and water conservation technologies.

 Mediating effect test

Based on the previous theoretical analysis, after experiencing the impact of MDS, farmers participate in collective actions, spread the risks and losses of MDS, and share the cost of post-disaster recovery to affect their behavior of adopting soil and water conservation technologies. In this section, stepwise regression, the Sobel test, and the Bootstrap test are employed to verify the mediating effect of collective action (see Tables 5–7 for details).

Collective action mediates the effect of MDS on engineering technology adoption behavior

Based on the stepwise regression test results of models (1), (7), and (8). Model (1) shows that MDS has a significant positive impact on farmers‘ engineering technology adoption behavior. Model (7) indicates that MDS also has a positive and significant effect on farmers' participation in collective action. Model (8) reveals that MDS has a positive but insignificant impact on farmers‘ engineering technology adoption behavior, while collective action exerts a significant positive effect on farmers' engineering technology adoption behavior. These results suggest that collective action plays a complete mediating role in the process where MDS affect farmers‘ engineering technology adoption behavior. Results from the Sobel test and Bootstrap test show the following. In the path where MDS affects farmers' engineering technology adoption behavior, the mediating effect value of collective action is 0.0528, accounting for 52.31% of the total effect. Its confidence interval is (0.0315, 0.0777), which is significant at the 95% confidence level (see Table 7 for details). Hypothesis 3 is thus confirmed.

 Collective action plays a mediating role in the impact of MDS on biotechnology adoption behavior

Based on the stepwise regression results of models (2), (7), and (9). In model (2), MDS has a significant positive impact on farmers' biotechnology adoption behavior. In model (7), MDS exerts a significant positive effect on farmers' participation in collective action. Model (9) shows that MDS has a significant positive impact on farmers' biotechnology adoption behavior, and collective action also has a significant positive impact on farmers' biotechnology adoption behavior. These results indicate that collective action plays a partial mediating role in the process where MDS affects farmers' biotechnology adoption behavior. It has a mediating role in the impact process (see Tables 5, 6). From the Sobel test and self-sampling test, it can be concluded that the mediation effect is 0.0232, accounting for 15.97% of the total effect, and the confidence interval is (0.0119, 0.0377), which is significant at the 95% confidence level. There was no masking effect (see Table 7). These results further confirm hypothesis 3.

Test of the mediating effect of collective action in the impact of MDS on farmers' adoption behavior of SWCT (Stepwise regression method result).

Variable name Collective action Engineering technology Biotechnology Farming technology
Model (7) Model (8) Model (9) Model (10)
MDS 0.3186*** (0.0407) 0.1598 (0.1345) 0.9921*** (0.2472) 0.1977 (0.1649)
Collective action 0.4994*** (0.0955) 0.5894*** (0.1483) 0.2878*** (0.1082)
Age −0.0001 (0.0013) 0.0031 (0.0041) 0.0006 (0.0054) 0.0052 (0.0046)
education level 0.0011 (0.0037) −0.0154 (0.0115) 0.0374** (0.0147) 0.0395*** (0.0126)
The proportion of male labor force 0.0138 (0.0733) 0.2358 (0.2294) 0.2504 (0.3011) 0.0167 (0.2548)
Land management area 0.0014 (0.0010) 0.0212*** (0.0035) 0.0135*** (0.0035) 0.0077*** (0.0029)
Village cadres 0.0551 (0.0476) −0.0949 (0.1496) 0.0411 (0.1814) 0.0237 (0.1545)
Number of farm machinery 0.0080 (0.0248) 0.0969 (0.0769) 0.1126 (0.0985) 0.3172*** (0.0799)
Proportion of non-farm income 0.0815* (0.0421) −0.0323 (0.1319) 0.2046 (0.1752) −0.0651 (0.1443)
Government subsidies 0.0921*** (0.0289) 0.2455*** (0.0894) 0.0965 (0.1240) −0.0138 (0.0976)
Social network 0.0003* (0.0002) 0.0001 (0.0005) −0.0006 (0.0006) −0.0008 (0.0006)
Technology ecological value cognition 0.0417** (0.0164) 0.0175 (0.0500) 0.2136*** (0.0609) 0.2008*** (0.0529)
Gansu (with Shaanxi as the reference group) −0.1146*** (0.0365) 0.1720 (0.1123) −1.2310*** (0.1561) 0.7241*** (0.1430)
Ningxia (with Shaanxi as the reference group) −0.0666* (0.0381) 1.0723*** (0.1171) −0.9177*** (0.1524) 1.0767*** (0.1470)
Constant term 0.2463** (0.1060) −0.8208*** (0.3046) −3.8341*** (0.4539) −1.7504*** (0.3352)
Prob > chi2 —— 0.0000 0.0000 0.0000
Log Likelihood —— −613.4169 −342.8219 −502.0662
Pseudo R 2 0.0916 0.1972 0.1853 0.1452
Wald test value 9.58*** 249.30*** 120.40*** 138.43***

*, ** and *** indicate significance at the statistical level of 10%, 5%, and 1%, respectively.

Test of the mediating effect of collective action in the impact of MDS on farmers' adoption behavior of SWCT (Sobel test and self-sampling test results).

Influence path Sobel test Self-sampling test (Bootstrap test) Is the mediation effect significant?
Indirect effect Direct effect Total effect Proportion of mediating effect or amount of masking effect Indirect effect 95% confidence interval
Lower limit Upper limit
i. MDS → collective action → engineering technology 0.0506*** (0.0120) 0.1371** (0.0447) 0.1877*** (0.0441) 52.31% 0.0506*** (0.0117) 0.0261 0.0789 The complete mediation effect is significant.
ii. MDS → Collective Action → Biotechnology 0.0326*** (0.0079) 0.0824** (0.0300) 0.1150*** (0.0295) 28.35% 0.0326*** (0.0071) 0.0205 0.0471 The mediating effect is significant.
iii. MDS → collective action → farming technology 0.0242** (0.0090) 0.1058** (0.0377) 0.1300*** (0.0369) 18.65% 0.0242** (0.0091) 0.0076 0.0439 The complete mediation effect is significant.

*, ** and *** indicate significance at the statistical level of 10%, 5%, and 1%, respectively.

 Collective action plays a mediating role in the impact of MDS on farming technology adoption behavior

Based on the stepwise regression results of models (3), (7), and (10): In model (3), MDS has a significant positive impact on farmers' farming technology adoption behavior. In model (7), MDS exerts a significant positive effect on farmers' participation in collective action. In model (10), the influence coefficient of MDS on farmers' farming technology adoption behavior is positive but insignificant, while collective action has a positive and significant impact on farmers' farming technology adoption behavior. These findings indicate that collective action plays a complete mediating role in the process where MDS affects farmers' farming technology adoption behavior (see Tables 5, 6). From the Sobel test and self-sampling test, it can be concluded that the mediation effect is 0.0250, accounting for 47.73% of the total effect, and the confidence interval is (0.0099, 0.0428), which is significant at the 95% confidence level. There was no masking effect (see Table 7). Hypothesis 3 was further verified.

 Robustness test

Since the explained variable is a binary selection variable, there may be a sample bias problem. In order to verify whether the impact of the above-mentioned MDS on farmers' SWCT adoption behavior is robust, this paper constructs a counterfactual analysis framework and uses the propensity score matching method (PSM) to measure the net effect of MDS on engineering technology, biotechnology, and farming technology in soil and water conservation technologies. Specifically, first, the experimental group was set as having MDS or participating in collective action and the control group as having no MDS and not participating in collective action. The logit model was used to calculate the score of each sample farmer. Second, three matching methods were selected for matching between the experimental group and the control group: K-nearest neighbor matching (K = 4), caliper matching (caliper = 0.020), and kernel matching (bandwidth = 0.060). Finally, the results of the average treatment effect (ATT) were obtained (as shown in Table 8). It is found that the results of the robustness test are basically consistent with the conclusions of the benchmark regression analysis.

Robustness test results of the impact of MDS and collective action on farmers' adoption behavior of SWCT.

Variable name Matching method ATT (MDS Experience) T-value Variable name Matching method ATT (Work in group) T-value
Engineering technology K-nearest neighbor matching (K = 4) 0.1032* (0.0563) 1.99 Engineering technology K-nearest neighbor matching (K = 4) 0.1797*** (0.0404) 5.03
Caliper matching (caliper = 0.020) 0.1124* (0.0519) 2.20 Caliper matching (caliper = 0.020) 0.1751*** (0.0346) 5.15
Kernel matching (broadband = 0.060) 0.1207** (0.0498) 2.46 Kernel matching (broadband = 0.060) 0.1751*** (0.0321) 5.26
Biotechnology K nearest neighbor matching (K = 4) 0.1332*** (0.0136) 6.36 Biotechnology K nearest neighbor matching (K = 4) 0.1055*** (0.0225) 5.55
Caliper matching (caliper = 0.020) 0.1331*** (0.0132) 6.43 Caliper matching (caliper = 0.020) 0.1057*** (0.0207) 5.74
Kernel matching (broadband = 0.060) 0.1332*** (0.0129) 6.64 Kernel matching (broadband = 0.060) 0.1306*** (0.0188) 5.73
Farming technology K nearest neighbor matching (K = 4) 0.0915* (0.0485) 2.50 Farming technology K nearest neighbor matching (K = 4) 0.0956** (0.0279) 3.35
Caliper matching (caliper = 0.020) 0.0736* (0.0436) 2.07 Caliper matching (caliper = 0.020) 0.1107*** (0.0253) 4.08
Kernel matching (broadband = 0.060) 0.0821* (0.0391) 2.48 Kernel matching (broadband = 0.060) 0.1109*** (0.0242) 4.16
Group difference analysis

The probit model was used above to measure the impact of MDS on the adoption of soil and water conservation technologies by farmers, but it could not reflect the differences in the adoption of soil and water conservation technologies by MDS of farmers with different characteristics. It is helpful to explore the group differences of different types of farmers to enrich the research content of MDS on farmers' technology choice behavior. based on previous research (Ademe et al., 2017), this paper divides the age of household heads into two categories: the young generation (55 years old and below) and the old generation (55 years old and above); the education level of household heads are divided into low education (9 years and below) and high education (more than 9 years) divide the scale of land management into two types—small-scale (less than 10 mu) and large-scale (10 mu and above). The Probit model was used to test the differences in the adoption behaviors of soil and water conservation technologies after different groups of farmers experienced MDS.

As shown in Table 9, there are obvious inter-group differences in the impact of MDS on farmers‘ SWCT adoption behavior, which shows that the age of the household head, the education level of the household head, and the scale of land management will affect the MDS of the farmers, and then affect farmers' adoption behavior of soil and water conservation technologies. In terms of different age groups, after the younger generation of farmers experienced MDS, the probability of adopting engineering technology increased by 40.37%; after the older generation of farmers experienced MDS, the probability of adopting farming technology increased by 54.20%. The younger the age, the more obvious the improvement in the adoption of engineering technology after experiencing MDS on the contrary, the older the household head, the greater the improvement in the adoption of farming technology after experiencing MDS. The possible reason is that the household head is young and can invest, the labor force is relatively large, and the adoption of engineering technology requires a large amount of continuous labor input, which leads to the fact that farmers with younger household heads tend to adopt engineering technology after experiencing MDS, the older the household head, the higher the degree of dependence on their own farmland.

Group difference analysis results of the impact of MDS and collective action on farmers' adoption behavior of SWCT.

Different age groups
Variable name Engineering technology Biotechnology Farming technology
Model (11) Model (12) Model (13) Model (14) Model (15) Model (16)
Young generation Old generation Young generation Old generation Young generation Old generation
MDS experience 0.4037** (0.1788) 0.2261 (0.1871) 1.4365*** (0.4512) 1.4365*** (0.4512) 0.2088 (0.2050) 0.5420** (0.2693)
Control variable Controlled Controlled Controlled Controlled Controlled Controlled
Constant term −1.9008*** (0.3286) −0.9661*** (0.3132) −3.3454*** (0.5968) −3.3454*** (0.5968) −2.8843*** (0.3988) −2.3098*** (0.4034)
Prob > chi2 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
Log Likelihood −345.0007 −273.7055 −179.5066 −167.9299 −303.9106 −192.9985
Pseudo R2 0.2009 0.1637 0.1801 0.172 0.0.1385 0.1637
Model (17) Model (18) Model (19) Model (20) Model (21) Model (22)
Lower education Highly educated Lower education Highly educated Lower education Highly educated
MDS experience 0.3078* (0.1607) 0.3518 (0.2267) 0.8727*** (0.2854) 1.7412*** (0.5031) 0.4036* (0.2220) 0.2086 (0.2399)
Control variable Controlled Controlled Controlled Controlled Controlled Controlled
Constant term −1.4993*** (0.3888) −1.9574*** (0.5937) −2.4896*** (0.5488) −3.0638*** (0.8662) −2.8221*** (0.4852) −2.5228*** (0.6281)
Prob > chi2 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
Log Likelihood −441.3698 −173.2128 −242.4249 −105.8656 −330.0077 −165.7577
Pseudo R2 0.1949 0.1969 0.1428 0.2280 0.1646 0.1141
Massive Small scale Massive Small scale Massive Small scale
MDS experience 0.2535 (0.1737) 0.4358** (0.1982) 1.1880*** (0.3559) 1.1341*** (0.3547) 0.3027 (0.2115) 0.3908 (0.2568)
Control variable Controlled Controlled Controlled Controlled Controlled Controlled
Constant term −0.8270* (0.4323) −1.6375*** (0.5109) −2.1787*** (0.6226) −3.3932*** (0.7669) −2.4861*** (0.4998) −3.5099*** (0.6941)
Prob > chi2 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
Log Likelihood −370.1559 −250.6521 −213.1237 −136.0230 −330.4184 −167.8957
Pseudo R2 0.1048 0.1612 0.1498 0.2004 0.1215 0.0983

As far as different educational levels are concerned, after the farmers with low education experience MDS, the probability of adopting engineering technology increases by 30.78%, and the probability of adopting farming technology increases by 40.36%. The improvement of SWCT's engineering technology adoption behavior and farming technology adoption behavior is more obvious. A possible reason is that compared with educated farmers, low-educated farmers lack re-employment ability after experiencing MDS and rely on land. If the degree is high, the enthusiasm for restoring agricultural production is strong, and they will tend to adopt farming technology, and they will also be more inclined to adopt engineering technology to increase the area of land management. As far as the scale of land management is concerned, after small-scale farmers MDS, the probability of adopting engineering technology increases significantly by 43.58%, while after large-scale farmers MDS, the improvement in engineering technology adoption is not significant, which fully demonstrates the driving effect of MDS on the adoption of engineering technology by small-scale farmers is that large-scale farmers tend to adopt engineering technology in order to avoid MDS and expand the scale of land management after experiencing MDS.

 Discussion

The discussion in this paper mainly includes the following two aspects: (1) Research findings and theoretical connection. This study verifies that MDS have a significant positive impact on farmers' adoption of SWCT, which is consistent with the logic of imprinting theory: disaster experiences drive farmers to take adaptive behaviors by shaping their risk perception and decision-making patterns. Meanwhile, the mediating effect of collective action indicates that farmers' cooperative behaviors (such as risk sharing and information sharing) after disaster shocks are an important transmission path for technology adoption, which echoes the view in Ostrom's collective action theory that “institutional arrangements promote the supply of public goods”. (2) Dialogue with existing research. Most existing studies analyze the impact of disaster shocks or collective action in isolation, while this study incorporates both into a unified framework, revealing the chain mechanism of “disaster shocks → collective action → technology adoption” and making up for the fragmented limitations of previous studies. For example, compared with studies that only focus on external factors such as government subsidies, this study emphasizes the role of farmers' endogenous cooperative behaviors, providing a more comprehensive explanation for understanding the promotion of SWCT, a technology with the attribute of public goods. It allows decision-makers, academics, and agricultural professionals to build focused policies that capitalize on the lessons acquired from previous catastrophes and harness the power of collaborative efforts. If we do this, we can make farming communities more resilient, reduce the adverse effects of climate change, and encourage sustainable agricultural practices that are good for farmers and the environment.

 Conclusion and policy implications Conclusion

This paper analyzed the impacts of MDS and collective action on farmers‘ adoption behavior of soil and water conservation technologies, using data from 1,106 respondents across Shaanxi, Gansu, and Ningxia provinces in China. The following conclusions are reached based on our analysis: First, among farmers' adoption behaviors of soil and water conservation technologies, engineering technologies account for 46.67% of the total sample, biological technologies account for 12.75%, and tillage technologies account for 22.33%. This indicates that the overall level of farmers‘ adoption of soil and water conservation technologies is not high, with particularly low adoption rates for tillage technologies and biological technologies. Second, MDS and collective action can significantly promote farmers' engineering technology adoption, biotechnology adoption, and farming technology adoption; third, collective action plays a role in the process of MDS on farmers‘ SWCT adoption. In terms of the positive mediation effect, there are significant inter-group differences in the impact of MDS on farmers' adoption of soil and water conservation technologies. Compared with other types of farmers, the improvement in behavior is more obvious; the farmers with household heads aged 55 years and be-low and with low education have experienced MDS, and the improvement of farming technology adoption behavior is more obvious.

 Policy implications

Based on the above conclusions, the following policy implications are drawn: (1) In response to the low adoption rates of tillage and biological technologies, it is necessary to optimize technology supply. Meanwhile, the “cooperative add demonstration household” model should be adopted to strengthen promotion efforts, thereby improving the popularity of these underdeveloped technologies. (2) Given the promotional effects of MDS and collective actions, a linkage mechanism integrating “risk perception-organizational collaboration-technology implementation” should be established. Specifically, early warning popularization and collective procurement can be utilized to reduce the costs of technology adoption. (3) Leveraging the mediating effect of collective actions, it is essential to improve institutional guarantees and interest linkage mechanisms. This will serve to enhance the pivotal role of collective actions in technology promotion. (4) Focusing on the key group of farmers whose household heads are aged 55 or below with low educational levels, hierarchical training and resource inclination policies should be implemented. Special subsidies and credit support can be employed to stimulate their enthusiasm for technology adoption, thus forming a precise and efficient promotion system.

 Data availability statement

The datasets presented in this article are not readily available because other participants are not allowed to make it public. Requests to access the datasets should be directed to dxl7482@gsupl.edu.cn.

 Ethics statement

The studies involving humans were approved by Gansu University of Political science and Law. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

 Author contributions

XD: Conceptualization, Data curation, Investigation, Methodology, Software, Validation, Writing – original draft, Writing – review & editing. LL: Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Validation, Writing – review & editing. QL: Conceptualization, Funding acquisition, Investigation, Project administration, Resources, Supervision, Visualization, Writing – review & editing. WZ: Methodology, Project administration, Resources, Supervision, Visualization, Writing – review & editing.

 Funding

The author(s) declare that financial support was received for the research and/or publication of this article. Youth Project of Humanities and Social Sciences, Ministry of Education; National Natural Science Foundation of China; Scientific Research Project of Gansu University of Political Science and Law Funding role: Provided relevant financial support for the core data collection, field research, preliminary analysis, academic exchanges, paper writing, and final improvement of the research results.

The authors would like to express their gratitude to the smallholder farmers who grow watermelons and muskmelons in the central and western regions of China for their time and participation during the field data collection. At the same time, the authors also want to thank those who have dedicated their time and efforts to this project and have sorted out relevant materials and data. These individuals include: Qu Yufei, Zhang Wenying, and many staff members and student assistants.

 Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

 Generative AI statement

The author(s) declare that no Gen AI was used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

 Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

 References Ademe Y. Kebede T. Mullatu A. Shafi T. (2017). Evaluation of the effectiveness of soil and water conservation practices on improving selected soil properties in Wonago district, Southern Ethiopia. J. Soil Sci. Environ. Manage. 8, 70–79. 10.5897/JSSEM2016.060138147025 Altobelli F. Vargas R. Corti G. Dazzi C. Montanarella L. Monteleone A. . (2020). Improving soil and water conservation and ecosystem services by sustainable soil management practices: from a global to an Italian soil partnership. Ital. J. Agron. 15:1765. 10.4081/ija.2020.1765 Ashoori D. Bagheri A. Sadegh M. Michailidis A. (2016). Understanding the attitudes and practices of paddy farmers for enhancing soil and water conservation in Northern Iran. Int. Soil Water Conserv. Res. 4, 260–266. 10.1016/j.iswcr.2016.09.003 Assefa S. Kessler A. Fleskens L. (2021). Using agent-based modelling to assess scenarios for enhanced soil and water conservation in the Boset District, Ethiopia. J. Artif. Soc. Soc. Simul. 24:1. 10.18564/jasss.4714 Avital E. Jablonka E. (1994). Social learning and the evolution of behaviour. Anim. Behav. 48, 1195–1199. 10.1006/anbe.1994.1351 Batjes N. H. (2014). Projected changes in soil organic carbon stocks upon adoption of recommended soil and water conservation practices in the Upper Tana River catchment, Kenya. Land Degrad. Dev. 25, 278–287. 10.1002/ldr.2141 Belachew A. Mekuria W. Nachimuthu K. (2020). Factors influencing adoption of soil and water conservation practices in the northwest Ethiopian highlands. Int. Soil Water Conserv. Res. 8, 80–89. 10.1016/j.iswcr.2020.01.005 Belayneh M. (2023). Factors affecting the adoption and effectiveness of soil and water conservation measures among small-holder rural farmers: the case of Gumara watershed. Resour. Conserv. Recycl. Adv. 18:200159. 10.1016/j.rcradv.2023.200159 Belayneh M. Yirgu T. Tsegaye D. (2019). Effects of soil and water conservation practices on soil physicochemical properties in Gumara watershed, Upper Blue Nile Basin, Ethiopia. Ecol. Process. 8, 1–14. 10.1186/s13717-019-0188-2 Bewket W. (2007). Soil and water conservation intervention with conventional technologies in northwestern highlands of Ethiopia: acceptance and adoption by farmers. Land Use Policy 24, 404–416. 10.1016/j.landusepol.2006.05.004 Bizoza A. R. (2014). Three-stage analysis of the adoption of soil and water conservation in the highlands of Rwanda. Land Degrad. Dev. 25, 360–372. 10.1002/ldr.2145 Brown P. Daigneault A. J. Tjernström E. Zou W. (2018). Natural disasters, social protection, and risk perceptions. World Dev. 104, 310–325. 10.1016/j.worlddev.2017.12.00230100664 Budhathoki N. K. Paton D. Lassa J. A. Zander K. K. (2020). Assessing farmers' preparedness to cope with the impacts of multiple climate change-related hazards in the Terai lowlands of Nepal. Int. J. Disaster Risk Reduct. 49:101656. 10.1016/j.ijdrr.2020.101656 Cao Y. Wang B. Guo H. Xiao H. Wei T. (2017). The effect of super absorbent polymers on soil and water conservation on the terraces of the loess plateau. Ecol. Eng. 102, 270–279. 10.1016/j.ecoleng.2017.02.043 Chen Y. Vanmaercke M. Jiao J. Bai L. Tang B. Wang N. . (2023). Quantifying the importance of different erosion processes and soil and water conservation measure collapses following an extreme rainstorm in the Chinese Loess Plateau. Land Degrad. Dev. 34, 403–422. 10.1002/ldr.4468 Chesterman N. S. Entwistle J. Chambers M. C. Liu H. C. Agrawal A. Brown D. G. (2019). The effects of trainings in soil and water conservation on farming practices, livelihoods, and land-use intensity in the Ethiopian highlands. Land Use Policy 87:104051. 10.1016/j.landusepol.2019.104051 Cui Z. Yu Z. Gong X. (2023). Study on the income-increasing effect of conservation tillage technology - A case study of reduced and no-tillage technology. Chinese J. Agric. Resour. Reg. Plan. 44, 59–69. Danso-Abbeam G. (2022). Do agricultural extension services promote adoption of soil and water conservation practices? Evidence from Northern Ghana. J. Agric. Food Res. 10:100381. 10.1016/j.jafr.2022.100381 Darkwah K. A. Kwawu J. D. Agyire-Tettey F. Sarpong D. B. (2019). Assessment of the determinants that influence the adoption of sustainable soil and water conservation practices in Techiman Municipality of Ghana. Int. Soil Water Conserv. Res. 7, 248–257. 10.1016/j.iswcr.2019.04.003 Dennis E. M. Brondizio E. (2020). Problem framing influences linkages among networks of collective action situations for water provision, wastewater, and water conservation in a metropolitan region. Int. J. Commons 14:18. 10.5334/ijc.974 Fallah M. Kavian A. Omidvar E. (2016). Watershed prioritization in order to implement soil and water conservation practices. Environ. Earth Sci. 75, 1–17. 10.1007/s12665-016-6035-1 Fontes F. P. (2020). SWCT adoption and labour allocation: evidence from Ethiopia. World Dev. 127:104754. 10.1016/j.worlddev.2019.104754 Fusco G. Vecchio Y. Porrini D. Adinolfi F. (2021). Improving the economic sustainability of Italian Farmer: an empirical analysis of decision-making models for insurance adoption. New Medit 3, 53–68. Gebremeskel G. Gebremicael T. G. Girmay A. (2018). Economic and environmental rehabilitation through soil and water conservation, the case of Tigray in northern Ethiopia. J. Arid Environ. 151, 113–124. 10.1016/j.jaridenv.2017.12.002 Gu C. Zhu Y. Li R. Yao H. Mu X. (2021). Effects of different soil and water conservation measures on hydrological extremes and flood processes in the Yanhe River, Loess Plateau, China. Nat. Hazards 109, 545–566. 10.1007/s11069-021-04848-w Hu Y. Zhang F. Luo Z. Badreldin N. Benoy G. Xing Z. (2023). Soil and water conservation effects of different types of vegetation cover on runoff and erosion driven by climate and underlying surface conditions. Catena 231:107347. 10.1016/j.catena.2023.107347 Huang X. Lu Q. Wang L. (2020a). Research on capital endowment, ecological cognition and farmers' water and soil conservation technology adoption behavior–based on the moderating effects of ecological compensation policies. Agric. Technol. Econ. 12, 33–44. Huang X. Lu Q. Wang L. Cui M. Yang F. (2020b). Does aging and off-farm employment hinder farmers' adoption behavior of SWCT in the Loess Plateau? Int. J. Clim. Change Strateg. Manag. 12, 92–107. 10.1108/IJCCSM-04-2019-0021 Huang X. Lu Q. Yang F. (2020c). The effects of farmers' adoption behavior of soil and water conservation measures on agricultural output. Int. J. Clim. Change Strateg. Manag. 12, 599–615. 10.1108/IJCCSM-02-2020-0014 Huang X. H. Wang L. L. Lu Q. (2019). Research on behavior of farmers' adoption of SWCT——an empirical analysis based on data collected from 1152 households on Loess Plateau. J. Northwest A&F Univ. (Soc. Sci. Ed.) 19, 133–141. Huo J. Yu X. Liu C. Chen L. Zheng W. Yang Y. Tang Z. (2020). Effects of soil and water conservation management and rainfall types on runoff and soil loss for a sloping area in North China. Land Degrad. Dev. 31, 2117–2130. 10.1002/ldr.3584 Indrawati D. R. Supangat A. B. Wahyuningrum N. Subandrio B. (2022). Community participation in soil and water conservation as a disaster mitigation effort. IOP Conf. Ser. Earth Environ. Sci. 1109, 012030. 10.1088/1755-1315/1109/1/012030 Jara-Rojas R. Bravo-Ureta B. E. Díaz J. (2012). Adoption of water conservation practices: a socioeconomic analysis of small-scale farmers in Central Chile. Agric. Syst. 110, 54–62. 10.1016/j.agsy.2012.03.008 Jarray F. Hermassi T. Mechergui M. Zucca C. Le Q. B. (2023). Long-term impact of soil and water conservation measures on soil erosion in a tunisian semi-arid watershed. Land 12:1537. 10.3390/land12081537 Jia R. Lu Q. (2018). Does land transfer promote the implementation of soil and water conservation measures by rural households in the Loess Plateau region? —— an analysis based on the mediating role of collective action and the moderating effect of government subsidies. Chin. Rural Econ. 6, 38–54. Jia R. Lu Q. (2019). The impact of migrant work and women's decision-making on farmers' participation in collective actions - A case study based on survey data of farmers in Shaanxi, Gansu and Ningxia Provinces. J. Agrotech. Econ. 122–134. Jiang C. Zhang H. Wang X. Feng Y. Labzovskii L. (2019). Challenging the land degradation in China's Loess Plateau: benefits, limitations, sustainability, and adaptive strategies of soil and water conservation. Ecol. Eng. 127, 135–150. 10.1016/j.ecoleng.2018.11.018 Kpadonou R. A. B. Owiyo T. Barbier B. Denton F. Rutabingwa F. Kiema A. (2017). Advancing climate-smart-agriculture in developing drylands: joint analysis of the adoption of multiple on-farm soil and water conservation technologies in West African Sahel. Land Use Policy 61, 196–207. 10.1016/j.landusepol.2016.10.050 Lei X. Li J. Li H. Yan J. Li P. Guo Y. . (2023). Regional assessment at the province level of agricultural science and technology development in China. Agriculture 13:389. 10.3390/agriculture13020389 Li L. Dingyi S. Xiaofang L. Zhide J. (2021). Influence of peasant household differentiation and risk perception on soil and water conservation tillage technology adoption-an analysis of moderating effects based on government subsidies. J. Clean. Prod. 288:125092. 10.1016/j.jclepro.2020.125092 Lin P. S. S. Chang K. M. (2020). Metamorphosis from local knowledge to involuted disaster knowledge for disaster governance in a landslide-prone tribal community in Taiwan. Int. J. Disaster Risk Reduct. 42:101339. 10.1016/j.ijdrr.2019.101339 Lingling L. Renzhi Z. Zhuzhu L. Weili L. Junhong X. Liqun C. Bellotti B. (2014). Evolution of soil and water conservation in rain-fed areas of China. Int. Soil Water Conserv. Res. 2, 78–90. 10.1016/S2095-6339(15)30015-032199392 Liu G. Dai F. Nan L. Du S. (2013). Assessing the suitability of soil and water conservation practices in Southwestern China. Land Use Policy 30, 319–327. 10.1016/j.landusepol.2012.03.022 Mao H. Chen S. Fu Y. (2024). Policy incentives and farmers' adoption of soil and water conservation technologies: an empirical analysis based on farmers in Shaanxi Province of the Yellow River Basin. J. Jiangxi Univ. Financ. Econ. 3, 91–101. 10.13676/j.cnki.cn36-1224/f.20231030.001 Marquis C. Tilcsik A. (2013). Imprinting: toward a multilevel theory. Acad. Manag. Ann. 7, 195–245. 10.5465/19416520.2013.766076 Mekuriaw A. Heinimann A. Zeleke G. Hurni H. (2018). Factors influencing the adoption of physical soil and water conservation practices in the Ethiopian highlands. Int. Soil Water Conserv. Res. 6, 23–30. 10.1016/j.iswcr.2017.12.006 Morris C. Arbuckle J. G. (2021). Conservation plans and soil and water conservation practice use: evidence from Iowa. J. Soil Water Conserv. 76, 457–471. 10.2489/jswc.2021.00166 Nigussie Z. Tsunekawa A. Haregeweyn N. Adgo E. Cochrane L. Floquet A. Abele S. (2018). Applying Ostrom's institutional analysis and development framework to soil and water conservation activities in north-western Ethiopia. Land Use Policy 71, 1–10. 10.1016/j.landusepol.2017.11.039 Nkegbe P. K. Shankar B. (2014). Adoption intensity of soil and water conservation practices by smallholders: evidence from Northern Ghana. Bio-based Appl. Econ. J. 3, 159–174. 10.13128/BAE-13246 Nyangena W. (2008). Social determinants of soil and water conservation in rural Kenya. Environ. Dev. Sustain. 10, 745–767. 10.1007/s10668-007-9083-612310376 Ogunniyi A. I. Omotayo A. O. Olagunju K. O. Motunrayo O. Awotide B. A. Mavrotas G. . (2023). Do soil and water conservation practices influence crop productivity and household welfare? Evidence from rural Nigeria. Environ. Sci. Pollut. Res. 30, 56016–56036. 10.1007/s11356-023-26280-236907915 Ojo T. O. Baiyegunhi L. J. Adetoro A. A. Ogundeji A. A. (2021). Adoption of SWCT and its effect on the productivity of smallholder rice farmers in Southwest Nigeria. Heliyon 7:e06433. 10.1016/j.heliyon.2021.e0643333763609 Ombogoh D. B. Tanui J. McMullin S. Muriuki J. Mowo J. (2018). Enhancing adaptation to climate variability in the East African highlands: a case for fostering collective action among smallholder farmers in Kenya and Uganda. Clim. Dev. 10, 61–72. 10.1080/17565529.2016.1174665 Ostrom E. Parks R. B. Percy S. L. Whitaker G. P. (1979). Evaluating police organization. Public Prod. Rev. 3, 3–27. Ostrom V. Ostrom E. (2019). “Public goods and public choices,” in Alternatives for Delivering Public Services (London: Routledge), 7–49. Pang J. Liu X. Huang Q. (2020). A new quality evaluation system of soil and water conservation for sustainable agricultural development. Agric. Water Manag. 240:106235. 10.1016/j.agwat.2020.106235 Parry M. L. (2019). Climate Change and World Agriculture. London: Routledge. Peng L. Tan J. Deng W. Liu Y. (2020). Understanding the resilience of different farming strategies in coping with geo-hazards: a case study in Chongqing, China. Int. J. Environ. Res. Public Health 17:1226. 10.3390/ijerph1704122632074990 Qin C. Tang Z. Chen J. Chen X. (2020). The impact of soil and water resource conservation on agricultural production-an analysis of the agricultural production performance in Zhejiang, China. Agric. Water Manag. 240:106268. 10.1016/j.agwat.2020.10626840215920 Qurniati R. Febryano I. G. Zulfiani D. (2017). How trust influence social capital to support collective action in agroforestry development? Biodiversitas J. Biol. Diversity 18, 1201–1206. 10.13057/biodiv/d180344 Rozaki Z. Wijaya O. Rahmawati N. Rahayu L. (2021). Farmers' disaster mitigation strategies in Indonesia. Rev. Agric. Sci. 9, 178–194. 10.7831/ras.9.0_178 Sileshi M. Kadigi R. Mutabazi K. Sieber S. (2019). Determinants for adoption of physical soil and water conservation measures by smallholder farmers in Ethiopia. Int. Soil Water Conserv. Res. 7, 354–361. 10.1016/j.iswcr.2019.08.00236711274 Streletskaya N. A. Bell S. D. Kecinski M. Li T. Banerjee S. Palm-Forster L. H. . (2020). Agricultural adoption and behavioral economics: bridging the gap. Appl. Econ. Perspect. Policy 42, 54–66. 10.1002/aepp.13006 Takayama T. Matsuda H. Nakatani T. (2018). The determinants of collective action in irrigation management systems: evidence from rural communities in Japan. Agric. Water Manag. 206, 113–123. 10.1016/j.agwat.2018.04.031 Teshome A. De Graaff J. Kassie M. (2016). Household-level determinants of soil and water conservation adoption phases: evidence from North-Western Ethiopian highlands. Environ. Manage. 57, 620–636. 10.1007/s00267-015-0635-526645077 Thapa P. S. Chaudhary S. Dasgupta P. (2022). Contribution of integrated watershed management (IWM) to disaster risk reduction and community development: lessons from Nepal. Int. J. Disaster Risk Reduct. 76:103029. 10.1016/j.ijdrr.2022.103029 Thornton P. K. Ericksen P. J. Herrero M. Challinor A. J. (2014). Climate variability and vulnerability to climate change: a review. Glob. Change Biol. 20, 3313–3328. 10.1111/gcb.1258124668802 Ullah R. Shivakoti G. P. Ali G. (2015). Factors effecting farmers' risk attitude and risk perceptions: the case of Khyber Pakhtunkhwa, Pakistan. Int. J. Disaster Risk Reduct. 13, 151–157. 10.1016/j.ijdrr.2015.05.005 Upadhaya S. Arbuckle J. G. Schulte L. A. (2023). Farmer typologies integrating latent and observed characteristics: insights for soil and water conservation outreach. Land Use Policy 134:106889. 10.1016/j.landusepol.2023.106889 VanOel P. R. Mulatu D. W. Odongo V. O. Willy D. K. Van der Veen A. (2019). Using data on social influence and collective action for parameterizing a geographically-explicit agent-based model for the diffusion of soil conservation efforts. Environ. Model. Assess. 24, 1–19. 10.1007/s10666-018-9638-y Villamayor-Tomas S. Thiel A. Amblard L. Zikos D. Blanco E. (2019). Diagnosing the role of the state for local collective action: types of action situations and policy instruments. Environ. Sci. Policy 97, 44–57. 10.1016/j.envsci.2019.03.009 Wang F. Mu X. Li R. Fleskens L. Stringer L. C. Ritsema C. J. (2015). Co-evolution of soil and water conservation policy and human–environment linkages in the Yellow River Basin since 1949. Sci. Total Environ. 508, 166–177. 10.1016/j.scitotenv.2014.11.05525478653 Wang J. Liu L. Zhao K. Wen Q. (2023). Farmers' adoption intentions of water-saving agriculture under the risks of frequent irrigation-induced landslides. Clim. Risk Manag. 39:100484. 10.1016/j.crm.2023.100484 Wang Y. Li X. Xu Q. Ying L. Lai C. Li A. (2022). Evaluation of soil and water conservation performance under promotion incentive based on stochastic frontier function model. Alex. Eng. J. 61, 855–862. 10.1016/j.aej.2021.04.095 Wang Y. Qian X. Zhou Y. Chen X. (2022). Spatial difference of Chinese public awareness of soil and water conservation and perception of information construction. Alex. Eng. J. 61, 8611–8623. 10.1016/j.aej.2022.01.048 Wei C. Dong X. Yu D. Liu J. Reta G. Zhao W. . (2022). An alternative to the Grain for Green Program for soil and water conservation in the upper Huaihe River basin, China. J. Hydrol. Reg. Stud. 43:101180. 10.1016/j.ejrh.2022.101180 Wen Z. Ye B. (2014). Analyses of mediating effects: the development of methods and models. Adv. Psychol. Sci. 22, 731–745. 10.3724/SP.J.1042.2014.0073137113526 Wigboldus S. Klerkx L. Leeuwis C. Schut M. Muilerman S. Jochemsen H. (2016). Systemic perspectives on scaling agricultural innovations. A review. Agron. Sustain. Dev. 36, 1–20. 10.1007/s13593-016-0380-z Wordofa M. G. Okoyo E. N. Erkalo E. (2020). Factors influencing adoption of improved structural soil and water conservation measures in Eastern Ethiopia. Environ. Syst. Res. 9, 1–11. 10.1186/s40068-020-00175-4 Xie Y. Sarkar A. Hossain M. S. Hasan A. K. Xia X. (2021). Determinants of farmers' confidence in agricultural production recovery during the early phases of the COVID-19 pandemic in China. Agriculture 11:1075. 10.3390/agriculture11111075 Yang S. Hundloe T. Cramb R. (2005). How China is adopting market mechanisms to foster soil and water conservation. Australas. J. Environ. Manage. 12, 37–46. 10.1080/14486563.2005.10648632 Zhang B. Tian L. He C. He X. (2023). Response of erosive precipitation to vegetation restoration and its effect on soil and water conservation over China's Loess Plateau. Water Resour. Res. 59:e2022WR033382. 10.1029/2022WR033382