Based on field survey data from 443 family farms in Shandong and Henan Provinces, the green control techniques(GCT) adoption behavior of family farms was measured in terms of adoption or non-adoption. Based on the global Moran’s I test, the Bayesian spatial Durbin probit model(BSDPM) was constructed, the appropriate spatial weight matrix was set, the optimal model for parameter estimation was selected, and the direct and spatial spillover effects of family farm characteristics on GCTs adoption behavior of family farms were decomposed by means of the partial differential method. The results show that the GCTs adoption behaviors of adjacent family farmers are spatially correlated and strongest when they are within 2.0 km of each other. Farm leaders’ educational level, degree of risk preference, financial status, number of laborers, understanding of GCTs and of the dangers of chemical pesticides, knowledge of other GCT adopters, frequency of communication with neighbors, participation in technical training and the strength of media publicity have significantly positive effects on the GCT adoption behaviors of family farms, which are mainly influenced by the direct effects of characteristic variables. However, the spatial spillover effects of neighboring family farmers’ participation in technical training, number of laborers, and financial status cannot be ignored. This result provides not only theoretical support for the demonstration and extension of the effectiveness of GCTs but also a reference for the selection of family farms as model households.

Determining an effective approach to replacing chemical fertilizer with organic fertilizer is a difficult challenge for the Chinese government. This paper constructs a dynamic analysis framework with broader application than previous statistical and case studies by theoretically deriving the impact of farmers' participation in contract farming on their organic fertilizer application behavior. The framework analyzes farmers' intertemporal organic fertilizer application behavior under the two scenarios of participation and nonparticipation in contract farming. Participation in contract farming positively impacts organic fertilizer application behavior in both the short and long terms. Survey data from 473 vegetable farmers in Shandong province of China were used to conduct an empirical analysis, and the endogenous switching probit model was used to solve the endogeneity problem of farmers' participation in contract farming. The empirical analysis supports the above results: farmers' participation in contract farming increases their probability of applying organic fertilizer by 50.7%. Robustness tests conducted using the recursive bivariate probit model and replacing the dependent variable (the intensity with which farmers replace chemical fertilizer with organic fertilizer) confirm the research results. Further heterogeneity analysis shows that farmers' participation in contract farming has a more obvious promoting effect on organic fertilizer application behavior in the older group, the group with an education level of primary school or below and the small-scale farming group. Therefore, the government should promote participation in contract farming, especially among elderly, low-education and small-scale farmers, to improve the adoption level of organic fertilizer in China.