In this paper, a unified framework of data extraction, variable coding, structural equation modeling and evaluation output is constructed around the evaluation of the mechanism of agricultural cooperatives in northern Anhui. The dataset integrates 84 cooperatives, 612 valid samples of farmers, questionnaires of farmers and village statistics from six prefecture-level cities in northern Anhui, and contains 21 observed variables and 4 potential dimensions. The study used expectation maximization to complete missing values, robust standardization of continuous variables, sequential coding of rank variables, and maximum likelihood estimation and 5000 Bootstrap sampling to complete parameter identification. The median value of comprehensive effectiveness of the six cities in northern Anhui was concentrated in the range of 65.0%-75.0%. The average values of comprehensive effectiveness, interest connection and participation response of comprehensive service cooperatives were 76.0%, 82.0% and 72.0%, respectively. The comprehensive explanation degree reached 67.0%.
