Determinants of Soil Conservation Adoption in Dry Lands of Saqqez County (An application of Generelized Ordered Logit Model)

Background and Objective: Soil protection is a prerequisite for realizing sustainable development and increasing the quantity and quality of food in countries. Soil erosion is one of the most important ecological and economic problems of agricultural and rural societies, which has significantly reduced the production of crops, also it Creates sediments in waterways and reducing performance in the water system, besides in causes inability of the soil to store the carbon cycle, nutrients and water, and is one of the biggest threats in the path of sustainability of agriculture. Soil erosion is one of the most important ecological and economic problems of agricultural and rural societies, which has significantly reduced the production of crops and the inability of the soil to store the carbon cycle, nutrients and water, and is one of the biggest threats in the path of sustainability. In order to prevent erosion and conserve soil, this study aims to determinants of soil conservation adoption among dry farmers in Saqqez County which can be used to develop strategies, policies, and appropriate programs for development, promotion and implementation of soil conservation technologies in effective dry farming.
 
Materials and Methods: This research centered on the community of dryland farmers in Saqqez County, Iran, encompassing a total population of about 7,099 individuals engaged in rainfed farming practices. To derive a representative sample, we applied the Cochran formula, which, under the assumptions of maximum variance (p = q = 0.5), a margin of error set at 5%, and a confidence level of 95%, yielded a sample size of 365 farmers. Data were gathered through a carefully designed questionnaire, distributed using proportionate stratified random sampling across the county's four administrative divisions: Central, Zivieh, Sarshiv, and Imam. This instrument explored the adoption patterns of seven specific soil conservation techniques—namely, crop rotation, application of manure, plowing perpendicular to slopes, retention of crop residues, fallowing, construction of stone bunds, and utilization of chemical fertilizers—alongside various socioeconomic variables. The dependent variable was conceptualized as an ordinal measure, reflecting the number of technologies adopted (ranging from 0 to 7), and further grouped into four distinct categories: no adoption (0), low adoption (1–2), moderate adoption (3–4), and high adoption (≥5). Initially, an ordered logit regression model was estimated to discern the factors influencing adoption. However, upon detecting a breach in the parallel regression assumption through the Brant test (χ² = 100.35, p < 0.01), we transitioned to a generalized ordered logit model, which better accommodates varying threshold parameters across different adoption thresholds. All estimations and analyses were executed using Stata version 14, employing maximum likelihood methods to obtain coefficients and assess model fit.
 
Results: According to the model, at zero level, the variable of the farmer’s record shows a negative sign and the variable of understanding the effects of the soil conservation expresses a positive sign. At level one, however, education variables, participation in soil conservation education classes, membership in rural institutions, area, non-agricultural employment and the number of parts are positive and the labor variable is negative. At the second level, education variables, participation in soil conservation education classes and dry stone maintenance are positive while the variables related to the experience of the farmer and non-agricultural employment show a negative significance.
 
Conclusion: According study’s results, Hence, government and regional authorities can adopt such policies as holding educational classes, providing farmers with facilities and creating non-agricultural jobs for them and take significant steps in conserving soil.
JEL Classifiation: Q2, Q5, Q24, Q26